Power Section Hummmm ....

[jbefumo]

My single-ended KT-120 amp sounded so good, I decided I should redesign the board to more neatly embody lessons learned along the way, adjust wire routing etc.  Now it has a loud hum with the volume turned all the way down.  I removed the power tube and jumped the signal to my little solid-state test amp, and there was no hum, so that seems to eliminate heater wiring in general and the preamp.


I jumpered the first (reservoir) capacitor (currently two 20u sections of a 20-20-40-40 can in parallel) to a big 500u/500V JJ, and the hum completely disappeared. Suspected maybe it was one/both of the capacitor sections, so I completely disconnected them.  Further experimentation suggests that a 50u capacitor leaves the hum.  the 500u makes it dead quiet.  However, the necessity of using such a large capacitor suggests that maybe it's just masking some more fundamental problem, and I'd far prefer to resolve the problem rather than hide the symptom.


I'll attach the schematic, though it's not 100% up to date.


Any ideas would be appreciated.

[sluckey]

I decided I should redesign the board to more neatly embody lessons learned along the way, adjust wire routing etc.  Now it has a loud hum with the volume turned all the way down.

Since it worked before you 'improved' it, I would examine the work you just did. Sometimes neat does not work.

That 500µf cap will likely kill that GZ34 tube. You may want to consider going with a full SS bridge. There is no benefit to using a tube rectifier in this amp.

[DummyLoad]

SE amps require a lot of filter. tube recto gains nothing in tone and puts limits on PS design. as you have already proven: more filter = quiet(er) operation. 


--pete

[jbefumo]

Yeah, just tried temporarily swapping in a solid state rectifier, just to eliminate that from the equation.  Still needs that 500u cap, so I'm going to go with that combination.  Good news is, I'm using (my last) stainless steel chassis, so eliminating the rectifier tube will allow me to mount the cap without having to punch another hole.


Trying to remember whether I made any other changes when I rebuilt it -- I seem to recall that I might have  [was still doing chemo, and my wits, such as they are, weren't quite up to par.]  Also, when I made my first video demo of it, I don't remember hearing the hum, but then heard one in the recording, and wrote it off as an artifact of the cheap vid recorder, but could well have been there all along, so ... Other good news is that the PT I'm using has a bunch of available taps, so I can recalibrate to get the voltage back where I want it,


Other than that, the amps sounds incredible.  Very clean and chimey, but that's what I was going for. 

I decided I should redesign the board to more neatly embody lessons learned along the way, adjust wire routing etc.  Now it has a loud hum with the volume turned all the way down.

Since it worked before you 'improved' it, I would examine the work you just did. Sometimes neat does not work.

That 500µf cap will likely kill that GZ34 tube. You may want to consider going with a full SS bridge. There is no benefit to using a tube rectifier in this amp.

[jbefumo]

Yup, thanks -- just reaching that conclusion.  Parenthetically, just built a true Class A1 push-pull amp for my brother -- an exercise in reusing spare parts that turned out surprisingly well.  It's based on the Hoffman Stout board, but I changed the tone control and used a pair of 6L6s making 18.5W, using values directly from an old 6L6 datasheet.  Will have to side-by-side them, but I'm reaching the conclusion that SE may be more trouble (and weight) than it's worth. Was actually just talking to Mercury about building me a transformer for a 45-50W single ended amp (2XKT88 in parallel), that I was going to use on the back end of an otherwise-Super-Reverb-Clone.  Still may ... just 'cause ... Interestingly, they had not problem doing it, and told me the package would be the same size and weight as the OT of a Marshall 100, which is less than I expected.


Oh, and THANKS (previous poster) for the tip about the rectifier tube -- was vaguely aware that there was some issue, but wasn't at all focused on that ....

SE amps require a lot of filter. tube recto gains nothing in tone and puts limits on PS design. as you have already proven: more filter = quiet(er) operation. 


--pete

[shooter]

(2XKT88 in parallel),

I *looked* down that road an she's ugly!
I've been building PSE's since I came on board here, the low watt's (<20)
they're fun, good sounding, an manageable, more than 20w, you're now living up to the SE *waste*, takes twice the money to get less than half the results.   There was a French? company that did offer a 40-50Watt tranny at about $400us and I believe it was about a pound a watt!

[DummyLoad]

hammond has a 75W offering of SE iron. the 1624SE is 5K @ 300mA and at a paltry 28 pounds.   

--pete

[jbefumo]

(2XKT88 in parallel),

I *looked* down that road an she's ugly!
I've been building PSE's since I came on board here, the low watt's (<20)
they're fun, good sounding, an manageable, more than 20w, you're now living up to the SE *waste*, takes twice the money to get less than half the results.   There was a French? company that did offer a 40-50Watt tranny at about $400us and I believe it was about a pound a watt!

I forget how much MM quoted me, but it wasn't much, if at all, more expensive than their other big transformers,  and the weight wasn't quite as bad at the Hammond unit I was looking at.  But yeah, getting 24-30W out of that KT120 took a bit of cogitation.  Tried hard to eliminate that high-voltage, low-current regulator chip, but just couldn't make it happen.  Pretty remarkable regulator, though.  Would have to use a pair of them for two KT-88s.  Guess it all comes down to the challenge, but then, if it were otherwise, I could have bought a wall of vintage amplifiers with what I spent getting to the point where I could build a single reliable one that sounds remotely decent....

[jbefumo]

I was actually looking at that, but I contacted them about the possibility of a minor change, and they had no interest, and did not recommend it as a musical instrument amp anyway.  I always dissed MM as 'boutique mojo" and smoke and mirrors, but I have to say, after several instances of contacting them about various possibilities, and not yet having bought a single thing (other than what came in a 'blackfaced' Super Reverb I recently acquired, their customer service is the best I've encountered anywhere.

hammond has a 75W offering of SE iron. the 1624SE is 5K @ 300mA and at a paltry 28 pounds.   

--pete

[Paul1453]

This statement:

Guess it all comes down to the challenge, but then, if it were otherwise, I could have bought a wall of vintage amplifiers with what I spent getting to the point where I could build a single reliable one that sounds remotely decent....

Comes from a man of experience and wisdom!      

[basschops1528]

...Tube recto gains nothing in tone and puts limits on PS design.

This might be going off track but thought tube rectos were sought after in SE amps, for instance the champ I'm going to build, to get some good sag and produce some good compression and overdrive?

[jbefumo]

That was my initial idea as well, but thinking about it, it would seem like the single-ended, design, or any true Class-A amp, is going to draw roughly the same current throughout the cycle, right?  I'm pretty sure that while a certain voltage drop is inherent in a particular rectifier tube, the sag/compression is a reaction to changes in current demand. I believe the issues I'm encountering are related to trying to wring the most possible power out of the tube in that configuration, rather than just being Class-A, but I'm sure someone more knowledgeable will correct me if I'm wrong.

...Tube recto gains nothing in tone and puts limits on PS design.

This might be going off track but thought tube rectos were sought after in SE amps, for instance the champ I'm going to build, to get some good sag and produce some good compression and overdrive?

[HotBluePlates]

...Tube recto gains nothing in tone and puts limits on PS design.

... thought tube rectos were sought after in SE amps, for instance the champ I'm going to build, to get some good sag and produce some good compression and overdrive?

Single-ended must be class A.

In class A, the output tube swings from idle current, to double-idle current, back to idle current, to zero current and back to idle. Average current stays essentially constant. There's a slight current increase for screen current, maybe preamps, but this is a very, very small percentage of the total current draw.

... to get some good sag and produce some good compression ...

Ohm's Law: Voltage [drop] = Current * Resistance

Since current stays essentially unchanged in class A, there is little-or-no change in voltage drop from idle to full power. So no sag from the rectifier. You can add sag in other ways, but don't count on big current in the B+ to get your sag/compression.

[jbefumo]

YEP YEP YEP! 


These facts were festering in the back of my so-called mind ... My initial approach was to use a high-voltage, low-current regulator chip for the screen supply, but that was giving me fits.  Would work fine for a while, then drop the voltage by 70V or so.  After messing with it for a while, one of the prior posts suddenly oozed into consciousness -- no big current changes -- no problem using a big honkin resistor.  Measured the current draw to the G2, applied the aforementioned Ohm's law, came up with a tentative value, tried a few variations -- bingo!  Works fine, sounds great, remains stable.  Sometimes the low-tech approach IS best...

...Tube recto gains nothing in tone and puts limits on PS design.

... thought tube rectos were sought after in SE amps, for instance the champ I'm going to build, to get some good sag and produce some good compression and overdrive?

Single-ended must be class A.

In class A, the output tube swings from idle current, to double-idle current, back to idle current, to zero current and back to idle. Average current stays essentially constant. There's a slight current increase for screen current, maybe preamps, but this is a very, very small percentage of the total current draw.

... to get some good sag and produce some good compression ...

Ohm's Law: Voltage [drop] = Current * Resistance

Since current stays essentially unchanged in class A, there is little-or-no change in voltage drop from idle to full power. So no sag from the rectifier. You can add sag in other ways, but don't count on big current in the B+ to get your sag/compression.

[HotBluePlates]

Going way back to the beginning of the thread...

You have hum, and suspect it's coming from the power supply/output tube. Power supply node A is the first filter cap after the rectifier, and also supplies the OT. That's fine in a push-pull amp (where hum applied to the OT CT cancels because it travels in opposite directions through the primary to the output tube plates), but not good in a single-ended amp.

Assuming you still have 16µF caps in parallel at node A, split them with a R or L to make a C-R-C or C-L-C filter. Attach the OT to the 2nd filter cap of this filter, not the cap connected directly to the rectifier. This will reduce the hum at the new node A, even if you have two 16µF filter sections instead of one 32µF section.

The L for the above will have to carrying the full current drawn through the B+ at all times. If you use an R, you will likely wind up with less filtering as well as more d.c. voltage drop.

[jbefumo]

Now have 500u/500V as the first cap, after replacing tube with solid state rectifier. B1.  Then the choke. Then 80u B2. 33k resistor goes to G2 .  10k resistor, then 40u and the preamp tubes.


So far, no smoke, no red plates, no sparks, sounds good.


Holding breath.

Going way back to the beginning of the thread...





You have hum, and suspect it's coming from the power supply/output tube. Power supply node A is the first filter cap after the rectifier, and also supplies the OT. That's fine in a push-pull amp (where hum applied to the OT CT cancels because it travels in opposite directions through the primary to the output tube plates), but not good in a single-ended amp.

Assuming you still have 16µF caps in parallel at node A, split them with a R or L to make a C-R-C or C-L-C filter. Attach the OT to the 2nd filter cap of this filter, not the cap connected directly to the rectifier. This will reduce the hum at the new node A, even if you have two 16µF filter sections instead of one 32µF section.

The L for the above will have to carrying the full current drawn through the B+ at all times. If you use an R, you will likely wind up with less filtering as well as more d.c. voltage drop.

[jbefumo]

Oh yeah, hum is gone.

Now have 500u/500V as the first cap, after replacing tube with solid state rectifier. B1.  Then the choke. Then 80u B2. 33k resistor goes to G2 .  10k resistor, then 40u and the preamp tubes.


So far, no smoke, no red plates, no sparks, sounds good.


Holding breath.

Going way back to the beginning of the thread...





You have hum, and suspect it's coming from the power supply/output tube. Power supply node A is the first filter cap after the rectifier, and also supplies the OT. That's fine in a push-pull amp (where hum applied to the OT CT cancels because it travels in opposite directions through the primary to the output tube plates), but not good in a single-ended amp.

Assuming you still have 16µF caps in parallel at node A, split them with a R or L to make a C-R-C or C-L-C filter. Attach the OT to the 2nd filter cap of this filter, not the cap connected directly to the rectifier. This will reduce the hum at the new node A, even if you have two 16µF filter sections instead of one 32µF section.

The L for the above will have to carrying the full current drawn through the B+ at all times. If you use an R, you will likely wind up with less filtering as well as more d.c. voltage drop.

[HotBluePlates]

Now have 500u/500V as the first cap, after replacing tube with solid state rectifier. B1.  Then the choke. Then 80u B2. 33k resistor goes to G2.

Oh yeah, hum is gone.

Gotcha; that's the brute-force method.

This same issue pops up all the time when folks here build Champ amps, but they use a good-quality 10- or 12-inch speaker which has decent low frequency response. A simple C-R-C or C-L-C filter feeding the OT fixes it every time even with small-value filter caps.

[jbefumo]

Sledge Hammers-R-Us ....


That's me -- strong as ox and nearly as smart ....

Now have 500u/500V as the first cap, after replacing tube with solid state rectifier. B1.  Then the choke. Then 80u B2. 33k resistor goes to G2.

Oh yeah, hum is gone.

Gotcha; that's the brute-force method.

This same issue pops up all the time when folks here build Champ amps, but they use a good-quality 10- or 12-inch speaker which has decent low frequency response. A simple C-R-C or C-L-C filter feeding the OT fixes it every time even with small-value filter caps.

[HotBluePlates]

Well, to be fair, the huge filter cap was not an option for others without switching to a solid-state rectifier. But most often, the builders already had higher-than-vintage supply voltage due to higher wall voltage being applied to an otherwise vintage-accurate PT. So even though the rectifier tube contributed no sag, they needed the voltage drop it afforded.

In those cases, the C-L-C was usually the most elegant & effective option.

If you didn't need to worry about output voltage, the big filter cap was the easiest approach.

[jbefumo]

As we used to say in the business -- make it work first, make it elegant after ....

Well, to be fair, the huge filter cap was not an option for others without switching to a solid-state rectifier. But most often, the builders already had higher-than-vintage supply voltage due to higher wall voltage being applied to an otherwise vintage-accurate PT. So even though the rectifier tube contributed no sag, they needed the voltage drop it afforded.

In those cases, the C-L-C was usually the most elegant & effective option.

If you didn't need to worry about output voltage, the big filter cap was the easiest approach.