ampage Tube Amps / Music Electronics	For current discussions, please visit Music Electronics Forum.

ampage archive

Vintage threads from the first ten years

rectifier diodes

9/8/1997 10:03 AM
Steve Keay	rectifier diodes In the Amp FAQS, R.G. has noted putting "1500V 1A diodes in series with the two sections of your rectifier tube" in case the tube shorts it can save your other components. Could someone clarify this? Not sure exactly what is meant by the two sections. Sorry if this is a dumb one. Also, Pitmans Tube book recommends installing 3 IN-4007 diodes "on each side of the output tubes that have the plate winding attached to them" which supposedly eats up transient spikes and increases tube life. Anybody shed some light on these mods? Are they needed? Why aren't these mods done by more manufacturers if they should be added? (except for the minimal cost reason) Thnaks in advance for info on these... Steve K

9/8/1997 10:03 AM

Steve Keay

rectifier diodes

In the Amp FAQS, R.G. has noted putting
"1500V 1A diodes in series with the two
sections of your rectifier tube" in case
the tube shorts it can save your other
components. Could someone clarify this?
Not sure exactly what is meant by the
two sections. Sorry if this is a dumb
one. Also, Pitmans Tube book recommends
installing 3 IN-4007 diodes "on each
side of the output tubes that have the
plate winding attached to them" which
supposedly eats up transient spikes and
increases tube life. Anybody shed some
light on these mods? Are they needed?
Why aren't these mods done by more
manufacturers if they should be added?
(except for the minimal cost reason)
Thnaks in advance for info on these...

Steve K

9/8/1997 11:47 AM
R.G.	Hi, Steve. Lemme see... A rectifier tube looks like two diodes, the anodes connected, one each, to the two wires from the power transformer, and the cathodes (bar end) connected together and also connected to the filter caps. The mod is to just put a silicon diode in series with each transformer wire, one per, into the rectifier tube, cathode to the anode of the rectifire tube. That way, the rectifier tube can short out and the diodes still prevent AC from getting to the filter caps and causing them to short out from reverse polarity, which would then kill the power transformer. The "3 - 1N4007's" are intended (I think...) to catch the spikes that are generated if the speaker lines ever open up or are connected to too high an impedance. The spike can cause arcing at the tube socket, burning a path across the socket and causing the socket to lead to tube runaway for that and successor tubes. It's probably simpler to put a 100 ohm 5W resistor across the output wires. The mods aren't used because the amps work withoug them, and they do cost money. Us DIY'ers, though, we can spend extra money on amps that already work....

9/8/1997 11:47 AM

R.G.

Hi, Steve. Lemme see...

A rectifier tube looks like two diodes, the anodes connected, one each, to the two wires from the power transformer, and the cathodes (bar end) connected together and also connected to the filter caps. The mod is to just put a silicon diode in series with each transformer wire, one per, into the rectifier tube, cathode to the anode of the rectifire tube.

That way, the rectifier tube can short out and the diodes still prevent AC from getting to the filter caps and causing them to short out from reverse polarity, which would then kill the power transformer.

The "3 - 1N4007's" are intended (I think...) to catch the spikes that are generated if the speaker lines ever open up or are connected to too high an impedance. The spike can cause arcing at the tube socket, burning a path across the socket and causing the socket to lead to tube runaway for that and successor tubes. It's probably simpler to put a 100 ohm 5W resistor across the output wires.

The mods aren't used because the amps work withoug them, and they do cost money. Us DIY'ers, though, we can spend extra money on amps that already work.... :-)

9/8/1997 12:20 PM
Steve Keay	Thanks for the info RG. And for all the info for that matter... I do remember hearing in the past (70's) from at least one Vox owner of power rectifier burnouts in AC30 amps so I want to try to address any issues in this early stage I'm at in the build process... Think I'll go with your suggestion for the rectifier diodes at least for now. Thanks again... Steve K

9/8/1997 5:37 PM
Charles	>>The mods aren't used because the amps work withoug them, and they do cost money. Us DIY'ers, though, we can spend extra money on amps that already work....<< I dunno... Fender has been putting 3kv diodes on the output primary halves for some time. I was amazed to find the same on the 50 watt and 60 watt cheapie Sovtek amps too. I'm not sure it a quantitative amount of help can be claimed tho. I've a question: Does putting in the diodes on the PT mess up the B+ voltage any? Or is it insignificant? Charles

9/8/1997 5:37 PM

Charles

>>The mods aren't used because the amps work withoug them, and they do cost money. Us DIY'ers, though, we can spend extra money on amps that already work....<<

I dunno... Fender has been putting 3kv diodes on the output primary halves for some time. I was amazed to find the same on the 50 watt and 60 watt cheapie Sovtek amps too. I'm not sure it a quantitative amount of help can be claimed tho.

I've a question: Does putting in the diodes on the PT mess up the B+ voltage any? Or is it insignificant?

Charles

9/9/1997 10:25 AM
Bruce	Charles... Bruce here again, Yes it does effect the power supply voltage.... but in a positive way! Yikes... no pun intented. The diodes drop the AC voltage ahead of the rectifier tube by an insignifcant amount..... The amount of the voltage drop across the SS diode junction. .6v .7v or so. Big deal. But it does give the whole power supply and filter section a leg up on the ripple effect and now the rectifier tube's anodes sees only positive going voltage too. I can see on my scope a difference in the ripple with the diodes ahead of the rectifier tube. Maybe it has something to do with the additioanl voltage drop the tube has and it acts like a nother filter section or something. I bet that makes life easier for the tube too. This little mod the RG is talking about is a must (IMO) for anyone using 6V6 tubes in a DIY amp. Those new 6V6's seem to blow up real easy with anything over 375V to 425V. I put the diodes in EVERY amp I fool around with because it is transparent in sound and assures that no AC can get to the filter caps and power tubes, .etc .etc. I also put a diode after the smoothing choke or first dropping resistor to keep the power amp from pulling current (under a big load) from the rest of the power supply caps. Sounds silly, but try it. That mod does make a little difference in sound, but I like what it does. ************************************** By the way RG, Have you read all the stories about bypassing the Hi voltage diodes with a hi ohm resistor and .01 cap...etc? 220K-470K .01 ?? I read a great article in one of my older ham radio rags about the futility of that and how it makes absolutley no differnce with modern SS diodes, execept to make sure the diodes see the same voltage divide across the resistors. But totaly unecessary and a waste of good parts. Hence the reason some prefer to use 3 diodes in series I think. I use 2 in each leg myself. But on BIG homebrew RF amps at 4000 V @ 1 amp I've used 6-8 diodes (each leg) in a voltage doubler to be safe. I used one 1n4007 diode in test jig of mine and tried like crazy blow the diode up at 800 VDC through a big huge 650ohm ceramic tube type resistor to ground and could not do it. Yes the tranny dropped a ton of voltage, but I still could measure close to 600 VDC on the resistor. That was close to 500 watts DC!! In one teeny SS diode! Remarkable! Why three? Probably the peak reverse voltage scare thing? another novel... Bruce

9/9/1997 10:25 AM

Bruce

Charles... Bruce here again,
Yes it does effect the power supply voltage....
but in a positive way! Yikes... no pun intented.
The diodes drop the AC voltage ahead of the
rectifier tube by an insignifcant amount.....
The amount of the voltage drop across the SS
diode junction. .6v .7v or so. Big deal.
But it does give the whole power supply and
filter section a leg up on the ripple effect and
now the rectifier tube's anodes sees only
positive going voltage too.
I can see on my scope a difference in the ripple
with the diodes ahead of the rectifier tube.
Maybe it has something to do with the additioanl
voltage drop the tube has and it acts like a nother
filter section or something.
I bet that makes life easier for the tube too.
This little mod the RG is talking about is a must
(IMO) for anyone using 6V6 tubes in a DIY amp.
Those new 6V6's seem to blow up real easy with
anything over 375V to 425V.
I put the diodes in EVERY amp I fool around with
because it is transparent in sound and assures
that no AC can get to the filter caps and power
tubes, .etc .etc.
I also put a diode after the smoothing choke or
first dropping resistor to keep the power amp from
pulling current (under a big load) from the rest
of the power supply caps. Sounds silly, but try it.

That mod does make a little difference in sound,
but I like what it does.
**************************************
By the way RG, Have you read all the stories about
bypassing the Hi voltage diodes with a hi ohm
resistor and .01 cap...etc? 220K-470K .01 ??
I read a great article in one of my older ham radio
rags about the futility of that and how it makes
absolutley no differnce with modern SS diodes,
execept to make sure the diodes see the same
voltage divide across the resistors. But totaly
unecessary and a waste of good parts.
Hence the reason some prefer to use 3 diodes in
series I think. I use 2 in each leg myself.
But on BIG homebrew RF amps at 4000 V @ 1 amp I've
used 6-8 diodes (each leg) in a voltage doubler
to be safe.
I used one 1n4007 diode in test jig of mine
and tried like crazy blow the diode up at 800 VDC
through a big huge 650ohm ceramic tube type resistor
to ground and could not do it. Yes the tranny
dropped a ton of voltage, but I still could measure
close to 600 VDC on the resistor. That was close
to 500 watts DC!! In one teeny SS diode! Remarkable!
Why three?
Probably the peak reverse voltage scare thing?

another novel...

Bruce

9/12/1997 4:54 PM
R.G.	Actually, there are two ways to do this, each with slightly different results. One is to put a single diode in series with each rectifier tube anode, the other is to connect the cathodes of the diodes to make a full wave rectified point, and then tie BOTH anodes of the rectifier tube to the cathodes of the SS rectifiers. In the first way, the tube rectifier acts as it normally would, alternate anodes conducting. In the second way (which is what I think you are talking about here) both anodes conduct every half cycle, effectively in parallel. The difference is that in the first case, the "plate resistance" of the tube rectifier is the same as it would be if there were no SS diodes there at all; in the second, since both halves of the rectifier tube are in parallel at all times, the effective internal resistance of the tube is halved. This will make some slight difference in the DC voltage and sound. The diode after the choke is a neat trick that stabilizes the power to the preamp stages; I saw that in the first issue of Glass Audio. On bypassing HV diodes, there are a couple of aims in doing that. In series strings, it does help to equalize the voltage distribution to keep from blowing the "weakest" diode first. In single diodes, though, it's a way of snubbing the turn off transient. As a diode turns off, it can reverse conduct for a very short time until all the carriers are swept out of the junction. This reverse conduction quits abruptly when all the carriers are gone, causing the diode to "snap off", a very abrupt change in current. This transient can literally be transmitted; it's a known source of EMI. The snubber helps that a lot. I've heard that the IR HexFRED diodes are very good for audio use because they have little or no snap off transients to get into the audio path.

9/12/1997 4:54 PM

R.G.

Actually, there are two ways to do this, each with slightly different results. One is to put a single diode in series with each rectifier tube anode, the other is to connect the cathodes of the diodes to make a full wave rectified point, and then tie BOTH anodes of the rectifier tube to the cathodes of the SS rectifiers.

In the first way, the tube rectifier acts as it normally would, alternate anodes conducting. In the second way (which is what I think you are talking about here) both anodes conduct every half cycle, effectively in parallel. The difference is that in the first case, the "plate resistance" of the tube rectifier is the same as it would be if there were no SS diodes there at all; in the second, since both halves of the rectifier tube are in parallel at all times, the effective internal resistance of the tube is halved.

This will make some slight difference in the DC voltage and sound.

The diode after the choke is a neat trick that stabilizes the power to the preamp stages; I saw that in the first issue of Glass Audio.

On bypassing HV diodes, there are a couple of aims in doing that. In series strings, it does help to equalize the voltage distribution to keep from blowing the "weakest" diode first. In single diodes, though, it's a way of snubbing the turn off transient. As a diode turns off, it can reverse conduct for a very short time until all the carriers are swept out of the junction. This reverse conduction quits abruptly when all the carriers are gone, causing the diode to "snap off", a very abrupt change in current. This transient can literally be transmitted; it's a known source of EMI. The snubber helps that a lot.

I've heard that the IR HexFRED diodes are very good for audio use because they have little or no snap off transients to get into the audio path.

9/12/1997 8:23 PM
Bruce	I connected up to Microsemi and downloaded the tech file on the HexFREDs a few weeks ago and I came to the same conclusion. The device I was checking out looks like a chassis mount microwave oven diode set. ie. 2 diodes tied at their cathodes. Looks ok for a small power supply and I wonder if it would add tonally to the amp becasue of those characteristics. BUT, what I was really looking at was using them as a clipper stage in series with a few Ge diodes (just to raise the clamping voltage up but keep the softer diode actio of the Ge diodes).. and see if it would give a nice roundy smooth tone to the clipping circuit. And also.... I looked over all the displays in reference to the diode clippers on DMZ. Why does you think the green LEDs in the clipping circuit pass so many even harmonics compared to the red, yellow LEDs or regular switching diodes ? I have even used IR diodes. Seems like all diodes have different sounds to some degree. I find that very interesting and worth looking into and playing with it some more. I used to play around with Hewlett Packard 1N2835 Schottky diodes in FWB (ring) balanced modulators and FWB (ring) product detectors circuits in radios. They always were quieter, smother and more round sounding for some reason. I wonder if there is any potential there. Thanks for all the great info RG! Bruce

9/12/1997 8:23 PM

Bruce

I connected up to Microsemi and downloaded the
tech file on the HexFREDs a few weeks ago and I
came to the same conclusion.
The device I was checking out looks like a chassis
mount microwave oven diode set. ie. 2 diodes
tied at their cathodes.
Looks ok for a small power supply and I wonder if
it would add tonally to the amp becasue of those
characteristics.
BUT, what I was really looking at was using them
as a clipper stage in series with a few Ge diodes
(just to raise the clamping voltage up but keep
the softer diode actio of the Ge diodes).. and
see if it would give a nice roundy smooth tone
to the clipping circuit.

And also....
I looked over all the displays in reference to
the diode clippers on DMZ. Why does you think
the green LEDs in the clipping circuit pass so
many even harmonics compared to the red, yellow
LEDs or regular switching diodes ?
I have even used IR diodes. Seems like all diodes
have different sounds to some degree.
I find that very interesting and worth looking
into and playing with it some more.
I used to play around with Hewlett Packard 1N2835
Schottky diodes in FWB (ring) balanced modulators and
FWB (ring) product detectors circuits in radios.
They always were quieter, smother and more
round sounding for some reason.
I wonder if there is any potential there.
Thanks for all the great info RG!
Bruce

Page 1 of 3

Next>

Last Page>>