ampage
Tube Amps / Music Electronics
For current discussions, please visit Music Electronics Forum. New: view Recent Searches.
New: visit Schematic Hell!
The sunn still shines online!

 
Listen to great tunes streaming live right now!

ampage archive

Vintage threads from the first ten years

Search for:  Mode:  
previous: Rick Erickson Re: Solen Caps in Power Supply -- 1134511980 View Thread

warning LONG!

12/13/2005 8:21 PM
Dai H.
email
warning LONG!
hi, check out this thread, plus the old thread it refers to:  
 
http://www.firebottle.com/fireforum/fireBB.cgi?cfg=ga&forum=gamt&thread=339448-000000.msg  
 
 
--------------------------------------------  
START NEW THREAD REPLY Top of Thread LIST  
Read 237 times  
 
From: Chuck @  
Date: 6/18/2005 8:36 PM  
Subject: Filter caps in series Q?  
 
When using two filter caps in series to ground, like you see on some amps main filter, do you double the ESR? Likewise, does putting two filters in parallel to ground half the ESR?  
 
Thanks  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 230 times  
 
From: Winnie Thomas @  
Date: 6/18/2005 8:59 PM  
Subject: Re: Filter caps in series Q?  
 
First of all, I don't know what ESR means.  
 
That being said, in series you add the capacitance inversely. For example, two 40uf caps is series yields a total of 20uf ( 1/40 + 1/40 = 2/40 invert 40/2 = 20). in series, the voltage adds. 40uf/350V in series with 40uf/350V- 20uf/700V.  
 
They sometimes have a 220K 1W resistor in parallel with each cap to assure that the voltage is distributed evenly between both caps.  
 
Look at the filter cap section of a Blackface Fender Bassman Schematic.  
 
Winnie  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 214 times  
 
From: Chuck @  
Date: 6/19/2005 4:38 AM  
Subject: Re: Filter caps in series Q?  
 
Thanks, I'm pretty good with all that though. I'm actually trying to get a grip on the effect of ESR on the circuit and figured this Q would be a place to start. ESR is "equivalent series resistance". All caps have some and bad ones have more. I've read that high ESR can cause some problems like ground loop hum and ghost notes. But I'm wondering how much of an issue it really is since it seems to me that series filter capacitors = series ESR. If so, then it hasn't been a problem for the amps that employ this in their design. So what is the real effect of ESR on a circuit?  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 202 times  
 
From: Greg Simon  
Date: 6/19/2005 7:38 AM  
Subject: Re: Filter caps in series Q?  
 
You know Chuck, I never really thought about this, but now that you mentioned it, I find that I'm really interested in the answer too. I've done a couple amps where I've used series filters, and didn't notice any hum difference, so it makes me wonder. Anyone have th answer to this?  
 
Greg  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 205 times  
 
From: Dai Hirokawa @  
Date: 6/19/2005 8:53 AM  
Subject: Re: Filter caps in series Q?  
 
probably if you're talking about regular alu electrolytics, it's something you can measure, but won't make much of a difference in function or sound. It's more when you replace an alu electro with something significantly different like a film cap, which should have lower ESR that you may have problems. Say, if the first one after the rectifier was changed to a film, and the faster response created problems for the PT or tube rectifier due to higher peak charging currents or whatever.  
 
Real caps aren't the "ideal cap"("C") that you see in formulas, right? They might be better or worse at higher frequencies, have differences in volumetric efficiency (capacitance provided vs. space taken), variance under temperatures, distortion, cost difference, etc. So a regular alu electro won't be acting as a cap at higher frequencies a film cap would and have higher impedance. The film would have better characteristics in the low freqs. too, so you should be getting lower impedance in the power supply (faster charge/discharge), and coupling frequencies better. Say, a regular 10uF alu electro might not be acting as a cap at 8Hz and 300kHz, (charging/storing, or coupling frequencies) while a 10uF film would (not genuine specs that I looked up but just a general idea!!!), or maybe to get similar low freqs., w/ a 100uF alu electro, you find you need to back down on the size when going to film and end up w/a 47uF or whatever. Look at the first pic in this link, and transpose it w/the 5F6-A first stage (change the output Xfrmer for the 100k plate load R) :  
 
http://www.nutshellhifi.com/library/ETF.html  
 
The AC (AC as in the output signal) plate current loop goes *through* the 100uF "filter/de-coupling cap" (which would be 8uF in the Bassman) as it goes from the plate back to the cathode of the triode. So it looks evident that the characteristics of that cap (AND the cathode bypass--which shorts the AC to the cathode) can influence the measured characteristics and sound. If you look in tube textbooks, at a "triode model", it shows the plate load grounded, because the power supply is at (AC) ground, and if you think about it, it's the (local) filter/de-coupling cap that is keeping it at (AC) ground.  
 
In the past, when I barely had a clue what I was doing, for example I changed the 220uF filter (regular cheap alu elctro) in a GCB-95 wah to a 220uF Black gate (fancy low ESR type marketed to audiophiles), and the wah effect didn't work until I backed off the uFs to 100uF or less, so (in hindsight) apparently it was the diff. in characteristics making the difference. So when you think about caps, it probably helps to think about the characteristics, frequency, not just the "uF" value, esp. if it could lead to trouble or damage. Another example, in Sanyo OS-con (organic semiconductive electrolytic caps--a very low ESR cap) literature, they say something about if you change to them on a circuit having ordinary al electros, the oscons can cause oscillation in a circuit that had oscillations damped from the higher ESR of the regular alu electros due to the lower ESR of the OS cons.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 188 times  
 
From: Wild Bill @  
Date: 6/19/2005 1:46 PM  
Subject: Re: Filter caps in series Q?  
 
I think sometimes we forget that we are comparing old specs/problems with new parts.  
 
ESR ratings today are FAR better than during the Golden Years! I doubt if the total ESR of two modern caps in series is as much as we had with one cap, in 1962.  
 
Also, ESR usually is not a problem with a new design but rather with an old cap coming to the end of its life cycle. ESR increases with age, sometimes dramatically.  
 
Film and oil-filled caps might be a better choice for lower ESR but they cost more and tend to be much bigger. An audiophile engineer might calculate he needs only a 10 mfd filter and chooses an expensive film filter, knowing it will have an acceptable low ESR. The fact that it's 10x the price is a marketing advantage in that world! A guy like Leo Fender would have chosen a far cheaper 22mfd or even a 47mfd electrolytic to ensure adequate de-coupling. His brute force solution was cheaper and faster than worrying about the ESR.  
 
.02  
 
---Wild Bill  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 175 times  
 
From: Dai Hirokawa @  
Date: 6/19/2005 3:37 PM  
Subject: Re: Filter caps in series Q?  
 
Also, ESR usually is not a problem with a new design but rather with an old cap coming to the end of its life cycle. ESR increases with age, sometimes dramatically.  
 
 
 
I think this is an interesting point, because maybe that provides some explanation for when people notice a change in tone w/new filters. Might be partly explained by higher ESR due to aging effects.  
 
 
Film and oil-filled caps might be a better choice for lower ESR but they cost more and tend to be much bigger. An audiophile engineer might calculate he needs only a 10 mfd filter and chooses an expensive film filter, knowing it will have an acceptable low ESR. The fact that it's 10x the price is a marketing advantage in that world! A guy like Leo Fender would have chosen a far cheaper 22mfd or even a 47mfd electrolytic to ensure adequate de-coupling. His brute force solution was cheaper and faster than worrying about the ESR.  
 
 
 
Another good point, it might be smarter(less expensive) to use a regular alu electro, but just increase the size.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 175 times  
 
From: Jack @  
Date: 6/19/2005 4:47 PM  
Subject: Re: Filter caps in series Q?  
 
Interesting comments, Bill... could this be some of the mojo of vintage amp tone?  
 
Anyhow, I was also curious about the series ESR thing, and some output stage ringing phenomena I observed on my Fender at frequencies around 5KHZ and up. I had fairly new Atoms in a standard 2-cap-in-series first (main) filter section and thought that if I lowered the power supply ESR, perhaps the ringing would reduce.... haven't tried that yet... but now I'm thinking that if I INCREASED the ESR, that actually might do the trick.  
 
Any ideas?  
 
Cheers,  
Jack  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 163 times  
 
From: Chuck @  
Date: 6/19/2005 7:35 PM  
Subject: Re: Filter caps in series Q?  
 
So what I'm gathering from all this is that the actual ESR as it relates to grounding AC from the power supply is nearly always low enough to be neglected. With the exception being grossly bad or aged caps. And that the actual power supply impedance is the real issue...Right?  
 
Thanks for the insight Bill  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 68 times  
 
From: Dai Hirokawa @  
Date: 6/22/2005 12:05 PM  
Subject: Re: Filter caps in series Q?  
 
Also, ESR usually is not a problem with a new design but rather with an old cap coming to the end of its life cycle. ESR increases with age, sometimes dramatically.  
 
 
 
was thinking about this again, specifically thinking of how in some layouts (tweed Fenders) the alu electros are right above the power tubes (or quite close to them or other hot parts of the layout) wondering if that contributes to accelerated aging compared to a chassis with a layout where they are located so they can't be heated as much. Wonder if that plays some small part in the difference between a "new sounding" amp vs. a "broken in" sounding amp. I would guess other things like the tubes losing highs over time, speakers breaking in would have a much bigger effect, but could aging of alu electros be a small part?  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 165 times  
 
From: Chuck @  
Date: 6/19/2005 7:29 PM  
Subject: Re: Filter caps in series Q?  
 
Woah!?! Thats more insight than normal amp guys should have. Especially in regard to intentionally designing inefficiency into a power supply so that the circuit design will work as it did with another less efficient filter. I suppose very low value resistors could be used at the filter grounds, with Solens for example, to get similar power supply characteristics and in some cases equivalent performance. Less can be more sometimes.  
 
But I always figured that since the filters main objective is to get the AC in the power supply to ground that the lowest possible ESR would be best. I'm still having trouble grasping any concept of the ESR coupling anything useful through the power supply.  
 
 
" If you look in tube textbooks, at a "triode model", it shows the plate load grounded, because the power supply is at (AC) ground, and if you think about it, it's the (local) filter/de-coupling cap that is keeping it at (AC) ground."  
 
Just like you said. So with the exception being the lowered power supply impedance effecting tube rectifiers there should only be benefits to lowest possible ESR. And you can always use smaller filters for rectifier tubes.  
 
It's alot to think about. And I'm sure any experiences I have like yours with the GCB-95 will raise a smile. (But only after prior events have raised some neck hair )  
 
By the way Dai, I always enjoy your thorough posting style.  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 161 times  
 
From: Jack @  
Date: 6/19/2005 7:47 PM  
Subject: Re: Filter caps in series Q?  
 
Chuck,  
 
Yep... good stuff! I would also tend to think that the closer you are to perfect AC ground in a power supply, the better... But in my case, I'm interested in the actual spurious "resonance" of the output stage, and how the ESR gets involved in that... there's an R-L-C thing happening there.  
 
I recall that in using simple snubber circuits, that an R-C combination works better than a straight cap.  
 
Jack  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 161 times  
 
From: Chuck @  
Date: 6/19/2005 8:02 PM  
Subject: Re: Filter caps in series Q?  
 
Perhaps as Dai mentioned about his GCB-95, there could be something more to the actual power supply impedance.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 157 times  
 
From: Wild Bill @  
Date: 6/20/2005 2:04 AM  
Subject: Re: Filter caps in series Q?  
 
Chuck, we should keep in mind that filters do more than just bypass AC ripple to ground. They actually provide the return path for the AC signal as well!  
 
The electrons travel from the heated cathodes to the plates, on to the load, over to the B+ node and through the filter to ground and back up to the cathode.  
 
If you follow the "other" convention of current flow the electrons travel exactly the same path, only in reverse.  
 
Whatever you prefer, the filters are in the electron path. If they have a high ESR or are too small there's an appreciable AC voltage drop across the cap. If the ESR to ground value from a B+ node is too high signals will couple to other stages on the same node rather than be bypassed to ground, always following the easiest path.  
 
Electrolytics have a higher impedance to ultrasonic frequencies. My failing old memory is telling me that's because the wet paste electrolyte can't respond as fast to follow the faster wave. This means the impedance portion of the ESR value would be higher at these frequencies.  
 
So this would be why some folks claim to hear a bit of high frequency "fizz" on some notes that is helped by using a film cap of a mfd or two across the last filter cap. The film adds only a trivial amount of extra capacitance to the electrolytic but it does a much better job of bypassing the "fizz" signal.  
 
I've never been able to hear any such "fizz" so I've never bothered to add a film but I don't deny that some folks truly hear it. I had a physics teacher long ago that got headaches from watching 16 mm films in class. His visual response time was quick enough that the picture to him had a bad flicker rate. I myself don't have much extra treble range to my hearing but my ears are much more sensitive to quiet signals than the rest of my family. (Years of wearing headphones and digging faint Morse code signals out of the noise level!)  
 
Different strokes for different folks...  
 
---Wild Bill  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 145 times  
 
From: Dai Hirokawa @  
Date: 6/20/2005 10:20 AM  
Subject: Re: Filter caps in series Q?  
 
Thanks Chuck.  
 
 
But I always figured that since the filters main objective is to get the AC in the power supply to ground that the lowest possible ESR would be best. I'm still having trouble grasping any concept of the ESR coupling anything useful through the power supply.  
 
 
 
 
Just like you said. So with the exception being the lowered power supply impedance effecting tube rectifiers there should only be benefits to lowest possible ESR. And you can always use smaller filters for rectifier tubes.  
 
 
 
Strictly from a technical performance perspective, the lower the ESR the better, but as filter/de-coupling caps they can affect the frequency response. And say, if you went to extremes and made the ESR lower and lower affecting BOTH highs and lows--increasing the ability of that stage to amplify a wider and wider bandwidth--that might cause or contribute to problems like oscillation or make the stage more susceptible to noise plus maybe other problems. Maybe if you are trying to amplify a tremendously wide freq. range from elephant communications to GHz frequency cell phone signals, the lowest possible ESR throughout the freq. range would be the best thing. However, we're dealing with audio, which is a narrower range, and gtr. amps--typically even narrower still--so the idea that lowest possible ESR always being the best probably doesn't apply for what we're engaged in. Restricting/limiting frequency ranges is not something that should automatically be taken as detrimental even though it can sound like you are taking something away or not letting it develop to it's full potential. Like say in a mix, it might be better to restrict the frequency range of different instruments so they don't get in the way of ea. other and make the sound muddy. An orchestra w/it's different instruments and freq. ranges. A gtr. amp w/lots of distortion--maybe it will sound much worse if you start extending the high freq. response.  
 
I guess this thread is what sparked these notions?  
 
http://www.firebottle.com/fireforum/fireBB.cgi?cfg=ga&forum=gadc&thread=324103-000000.msg  
 
Like you can see, it is possible to try raising the impedance up by using series resistance and inductance, which could solve the problems. (Also lower ESR can cause grounding probs but I haven't quite gotten my head around that.) And don't forget to try "normal" stuff like filtering, using caps to shunt highs to ground. Seems to be some sort of myth has developed that things like that will "kill the sound", but I think it depends so it wouldn't be wise to completely discount simple hi-cut filtering.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 99 times  
 
From: MBSetzer @  
Date: 6/21/2005 5:52 PM  
Subject: Re: Filter caps in series Q?  
 
*Restricting/limiting frequency ranges is not something that should automatically be taken as detrimental even though it can sound like you are taking something away or not letting it develop to it's full potential.*  
 
In RDH4 there are references to a comprehensive (at the time) study where human beings tended to tolerate unpleasant distortion better when the frequency response was narrower.  
 
Mike  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 164 times  
 
From: R.G.  
Date: 6/20/2005 12:31 PM  
Subject: Re: Filter caps in series Q?  
 
It's more when you replace an alu electro with something significantly different like a film cap, which should have lower ESR that you may have problems.  
 
 
Maybe, usually not. More as we go along.  
 
if the first one after the rectifier was changed to a film, and the faster response created problems for the PT or tube rectifier due to higher peak charging currents or whatever.  
 
 
This one's not a problem. The base C overwhelms the role of the ESR, except where damping rectifier shutoff ringing is concerned, and that's only a problem with solid state rectifiers that can stand the high current pulses.  
 
 
Real caps aren't the "ideal cap"("C") that you see in formulas, right?  
 
 
That part is correct. Real caps act like perfect caps with added parasitic parts. The reason we use the term ESR (Equivalent Series Resistance) is that the cap has the same response as a perfect cap in series with a small resistor. There are meters to measure ESR, by the way. EXTREMELY useful meters.  
 
There is also an ESL (... Inductance) that accounts for the inevitable rise in impedance at high frequencies, and a parallel leakage resistance taken to be the voltage across the cap divided by the leakage. You can do a bang up job of modelling the real cap by simulating the real cap capacitance with measured values for the other parasitic junk.  
 
All caps have a "bathtub" curve of impedance. At low frequencies they have a high impedance and drop off and 6db per octave like the stock Z = 1/(2*pi*F*C) equation predicts. At a frequency where the capacitive Z equals the ESR, they flatten out and the impedance is equal to the ESR until the rising Z = 2*pi*F*L of the ESL kicks in and the impedance now rises at 6db per octave. Note that even a perfect capacitor with ESR and ESL equal to 0 will have this same ESL rise because at some frequency the inductance of any leads attached to it becomes significant.  
 
Aluminum electros being wound in spirals where the current has to flow in what is effectively a coil have relatively high ESR and ESL because of the path length and shape. But they have high capacitance per unit and low cost, so they're the weapon of choice until frequency dictates that they just won't work.  
 
To say where an aluminum electro will or won't work, you cannot talk in generalities. You have to know the C, the ESR, and the ESL. High capacitance is pure aluminum electro territory. You can't get the big capacitance any other way. If you must have big capacitance and low ESR and ESL, you parallel numbers of modest sized caps or use specially wound and tabbed-out big ones.  
 
Notice that as aluminum electros get smaller, the ESR and ESL stay fairly small because the device is smaller too.  
 
Bypassing AL caps with films and ceramic is one way to restore the decreasing Z with frequency. When the AL flattens out you're fine, but when the ESL makes the impedance rise, you can put in a film or ceramic and get a decreasing Z with frequency until THAT cap's ESR and ESL turnovers are reached. It helps sometimes, although the hifi tweakos make too much of it.  
 
 
The film would have better characteristics in the low freqs.  
 
 
That's not necessarily true. At frequencies below the ESR/C turnover point, the capacitance of *any* cap dominates. So to make a film a "better" cap at frquencies below the ESR/C turnover point of the aluminum one, you'd have to have the ESR it self contributing a measurable effect. This happens essentially only with high frequency pulse/power use. Switching power supplies are the primary place.  
 
Sure the film has a lower ESR, but it has a lower capacitance too, so unless you're prepared to pay a lot for a big heavy cap that matches the aluminum's capacitance, you give away in capacitive impedance what you get in lower ESR. The devil is in the details. You have to do the numbers.  
 
Say, a regular 10uF alu electro might not be acting as a cap at 8Hz and 300kHz, (charging/storing, or coupling frequencies) while a 10uF film would (not genuine specs that I looked up but just a general idea!!!), or maybe to get similar low freqs., w/ a 100uF alu electro, you find you need to back down on the size when going to film and end up w/a 47uF or whatever.  
 
 
Case in point. From the datasheet at http://www.vishay.com/docs/28310/037rsm.pdf, they don't give you the ESR directly, they do show you the impedance versus frequency chart for some values and quote you the impedance at 10kHz.  
 
Interestingly, the worst Z versus F is the biggest cap, a 4700uF unit. It *starts* departing from ideal cap territory at a little over 1kHz. The caps get more "perfect" in that they look more like perfect caps to higher frequencies as the capacitance gets smaller. The 100uF on that chart looks like a cap out to over 4kHz. So you're not going to have a 10uF cap not look like a cap at 8 hz. It's going to look like capacitor like at over 10kHz, extrapolating from the chart.  
 
The devil is in the details. You gotta look up the numbers and calculate.  
 
 
The AC (AC as in the output signal) plate current loop goes *through* the 100uF "filter/de-coupling cap" (which would be 8uF in the Bassman) as it goes from the plate back to the cathode of the triode. So it looks evident that the characteristics of that cap (AND the cathode bypass--which shorts the AC to the cathode) can influence the measured characteristics and sound.  
 
 
Yes, they can - if the cap is bad. In the area where ESR << Zc, the cap looks like a cap. Gotta do the numbers.  
 
 
If you look in tube textbooks, at a "triode model", it shows the plate load grounded, because the power supply is at (AC) ground, and if you think about it, it's the (local) filter/de-coupling cap that is keeping it at (AC) ground.  
 
 
True. Capcitive impedance shows up as an impedance between the indicated circuit ground and the real ground. All stages share this impedance, so it looks like a feedback impedance. That's what causes it to have an effect on tone. Star grounding and individual stage decoupling can reduce this effect to almost being eliminated.  
 
 
 
In the past, when I barely had a clue what I was doing, for example I changed the 220uF filter (regular cheap alu elctro) in a GCB-95 wah to a 220uF Black gate (fancy low ESR type marketed to audiophiles), and the wah effect didn't work until I backed off the uFs to 100uF or less, so (in hindsight) apparently it was the diff. in characteristics making the difference.  
 
 
I'm kind of familiar with that circuit. That cap is there to filter the power supply, and early units didn't have it at all. I've used caps as high as 1000uF there with no ill effect. I suspect that you may have had some wiring/soldering problem when you put the 220uF cap in that was corrected when you took it out. It's easy to do stuff like that.  
 
 
So when you think about caps, it probably helps to think about the characteristics, frequency, not just the "uF" value,  
 
 
That's very true - but you have to do the numbers. Just whapping in film caps of lower value won't necessarily help, nor will bypassing AL caps with film by rote.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 151 times  
 
From: Dai Hirokawa @  
Date: 6/20/2005 1:57 PM  
Subject: Re: Filter caps in series Q?  
 
thanks R.G.! Lots to digest...  
 
 
I'm kind of familiar with that circuit. That cap is there to filter the power supply, and early units didn't have it at all. I've used caps as high as 1000uF there with no ill effect. I suspect that you may have had some wiring/soldering problem when you put the 220uF cap in that was corrected when you took it out. It's easy to do stuff like that.  
 
 
 
possibly, I'm not Harry Joyce but I think I'm pretty decent at soldering, and it's kind of hard to miss since it's a simple PC board. (Maybe it has something to do w/what I experienced? or not, but) when you see added power supply filtering in that circuit, it seems to always have a series R about 1k-1.5k. And maybe it's a false recollection, but I may have tried the filter cap only w/no series R and the wah didn't work? so I wonder if there is some need for the power supply to be higher impedance?  
 
 
Just whapping in film caps of lower value won't necessarily help, nor will bypassing AL caps with film by rote.  
 
 
 
right, and as one example IIRC, Paul Ruby experimented w/bypassing all filters w/0.1uF films and reported worse sound, so I guess it all depends.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 133 times  
 
From: Wild Bill @  
Date: 6/20/2005 7:04 PM  
Subject: Re: Filter caps in series Q?  
 
Thought I'd add another .02!  
 
I'm in the middle of repairing an old Gibson amp. I'd say it was made in the late 50's or so.  
 
Anyhow, the filters are bad of course. I've got 2 Planet brand "Lyticap" filters in my hand, one 20 mfd and one 16 mfd @ 450 vdc.  
 
In plain print it also says "Guaranteed for one year".  
 
I'll remember the next time we see one of those "I just found a 30 year old amp. Maybe I should replace the filters?" type posts.  
 
---Wild Bill  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 126 times  
 
From: Chuck @  
Date: 6/21/2005 2:15 AM  
Subject: Re: Filter caps in series Q?  
 
You know, it's amazing how long those things CAN last. I always replace filters in anything over 10 years old as a matter of course, but I've seen ALOT of old amps that sound fine, even great with the original filters in them.  
 
I got a Traynor Bassmaster head to use as my own personal mod platform about 5 years ago. It was all original and sounded GREAT stock (but that whole "stock" thing didn't last). And the filters were 21 years old!  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 101 times  
 
From: Greg Simon  
Date: 6/21/2005 6:52 PM  
Subject: Re: Filter caps in series Q?  
 
They put those Planet caps in the old Silvertones too and my brother's were still working ok in his 1484 head over 40 years after they were made. Amazing! But it does sound MUCH better with new caps.  
greg  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 109 times  
 
From: R.G.  
Date: 6/22/2005 4:24 AM  
Subject: Re: Filter caps in series Q?  
 
my brother's were still working ok in his 1484 head over 40 years after they were made. Amazing! But it does sound MUCH better with new caps.  
 
 
Yep. They keep on working... at some level.  
 
I prefer the way things sound when the caps are good. Some people don't.  
 
By the way, if you recap an amp for some non-tech but tone-picky diehard, you can "degrade" the new caps for them by inserting low value resistors in series as an external "ESR" to make it as bad as it used to be, so it still sounds "vintage".  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 101 times  
 
From: Greg Simon  
Date: 6/22/2005 5:32 AM  
Subject: Re: Filter caps in series Q?  
 
Thats a cool trick! I prefer the sound when the caps are new/good too, but I'll keep that in mind for the future!  
 
Greg  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 62 times  
 
From: Chuck @  
Date: 6/23/2005 9:59 PM  
Subject: THANKS TO ALL !!!  
 
I've been doing this as a hobbyist/hack for 16 years and after this thread I'm only now putting together the notion of the filters also being in the AC path. I love it when I learn something that has kept a piece of the circuit shrowded in mystery. So thanks a bunch  
 
Chuck  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 54 times  
 
From: Jack @  
Date: 6/24/2005 8:19 PM  
Subject: Re: Filter caps in series Q?  
 
Hi R.G.,  
 
Took a quick look at my Allied catalog, and checked out a few data sheets for commonly-used tube amp ATOM caps at Mouser, and don't seem to be able to find much info on ESR, at first glance anyhow.  
 
The only ESR posted values I saw were computer-grade caps and such, and they were rated in m-ohms (milli-ohms I assume), down around 100 mohms and such. So... I don't know how one is supposed to know what the ESR of any given amp cap is...?  
 
What resistance value range might you recommend for adding an external ESR to do a hypothetical cap vintage "downgrade" ?  
 
Cheers,  
Jack  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 58 times  
 
From: R.G.  
Date: 6/24/2005 11:57 PM  
Subject: Re: Filter caps in series Q?  
 
Yep - sometimes a direct spec of ESR is hard to find. Makers really don't want to be testing that, so they're goosy about specifying it.  
 
ESR is one of those things that is usually not needed in today's low voltage high current supplies. The frequency is too low. I suspect that if you look at capacitors specified for high frequency switching power supply outputs they may spec ESR.  
 
Here's what to do about ESR: Go get an ESR meter kit. There's an outfit in Australia (Oz to you that live there 8-) )that sells the kit, about US$80. It reads ESR directly. Dick Smith, as I remember.  
 
You can repair a lot of equipment by reading the ESR on the caps and replacing the suspicious ones.  
 
As to how much resistance is enough, the only good way to do that is to measure the ESR of the old, bad cap, then stick in a resistor of that value.  
 
They're usually in the low ohms range, under 100 for certain. OK, I found the manual.  
 
Yep, the ESR for a good cap is always under 20 ohms, and that's for a 1uF cap at 250V rating. There's a table of ESR values for good caps of various sizes and voltages.  
 
The manual says that ESR typically needs to get to 10-30 times the "good" value to cause trouble. Obviously, it can go as high as infinitely more than the normal (in an open cap) but somewhere in the 10-50 range is going to cause problems. So an old, bummer 22uF/250V cap that's about to die might have an ESR of 90 ohms. I'd measure the ESR, you could experiment with 5-50 ohms.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 53 times  
 
From: Jack @  
Date: 6/25/2005 4:53 AM  
Subject: Re: Filter caps in series Q?  
 
Good stuff, R.G. - great when you can actually put a number or range to something like this for future reference!  
 
One more silly question.... couldn't I save the $80. tab by just dialing my sine signal generator to 120 HZ, and putting a 600 ohm resistor in series with the cap in question, across the output, to determine ESR?  
 
Measuring the relative voltage drops across the 600R and the cap should tell you the impedance of the cap, right?  
 
Jack  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 65 times  
 
From: Wild Bill @  
Date: 6/25/2005 1:22 PM  
Subject: Re: Filter caps in series Q?  
 
You've got the right idea Jack. There's a bit more to it but if you go here:  
 
http://octopus.freeyellow.com/99.html  
 
you'll find what you need!  
 
---Wild Bill  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 56 times  
 
From: R.G.  
Date: 6/25/2005 1:33 PM  
Subject: Re: Filter caps in series Q?  
 
Yep, that's the measure-it-with-a-scope reference!  
 
Good searching.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 65 times  
 
From: R.G.  
Date: 6/25/2005 1:32 PM  
Subject: Re: Filter caps in series Q?  
 
One more silly question.... couldn't I save the $80. tab by just dialing my sine signal generator to 120 HZ, and putting a 600 ohm resistor in series with the cap in question, across the output, to determine ESR?  
 
 
No and maybe.  
 
The problem is that you have to separate the resistance part from the capcitance part. At 120Hz, the cap part dominates the impedance and so you can't measure the resistive part very well. That's the "no" answer.  
 
The "maybe" answer is: if you use a quite high frequency where the capacitive part of the ESR/cap combined impedance is tiny, all that's left is the resistor part. So put 100kHz on the 600R + cap. What's left across the cap is ESR+ESL. That is, by the way, what the ESR meters do in one form or another.  
 
There was one scheme I saw when I was researching ESR meters where you put the cap in a setup where you zap it with 100kHz square waves from a current source generator and watch the voltage waveform across the cap with a scope probe and measure the residual square waves across the cap. The height of the square waves is proportional to the ESR. But again, that's what the ESR meters do.  
 
It depends on how much you value your time and the accuracy of the answer I guess. When I first stumbled onto ESR meters I researched them on the net, and ran into a few schemos. I have one that I did a PCB layout for and was thinking of putting up at GEOFEX.  
 
I measured the ESR of exactly two caps by the hook-or-crook methods, messed with adjusting the scaling and stability of the crude analog meter for half a dozen more, then ordered the kit.  
 
I realize I forgot to mention - the ESR meters measure ESR in circuit. You don't have to remove them to see if a cap is bad (or middle-ish). You can debug an amp by first just whapping the ESR meter on every electro in the amp, just where they sit. When you've measured the dozen or so electro caps for ESR, you've eliminated or confirmed the idea that an electro cap is the problem. Takes about five minutes tops.  
 
You can do much the same thing with LCR bridges but they're usually more expensive than an ESR meter and don't work in circuit.  
 
START NEW THREAD REPLY PREVIOUS LIST  
Read 29 times  
 
From: Jack @  
Date: 6/29/2005 6:12 PM  
Subject: Re: Filter caps in series Q?  
 
Thanks for the input, RG and Bill!  
 
The fact that it can read ESR in-circuit is very cool !... that's quite a benefit to troubleshooting.  
 
Thanks,  
Jack  
 
----------------------------------------------------------------------------------

 
Replies:
Rick Erickson [QUOTE]I suppose very low value res... -- 12/14/2005 7:14 PM