Thanks Martin. I am adding the schematic for future referece. Will go through it in detail and see what I can come up with..
A quick question about Pentode operation. The VOX AC30 (with EF86) has the 100nF (C12) cap to ground, but when I was studying Pentode operation at valve wizard http://www.valvewizard.co.uk/pentode.html that cap (cg2)seems to go to the cathode of the tube...Which is the correct?
As far as the pentode I was reading this article today from tube-town and it seems more in line with valve wizard then the VOX regarding the screen bypass cap placement at cathode rather than ground (pic attached). https://www.tube-town.net/cms/?Info/How_To_Use_a_EF86
Diodes are reversed and centre tap is not needed; no ground connection there. Search for full wave voltage doubler. Here for example http://www.tubecad.com/january2000/page14.html You also won’t need high voltage caps, I think 2x 47u/160v will be enough for the doubler, 250v for the extra filtering. Keeps it small and cheaper
On the extra filtering part: seems like a good idea. You do not need -200V for full cutoff. I’d advise against just regulating the negative supply if you take the bias from it. That would make it resistant against line voltage variations. Screen and plate voltage won’t be though, so a line voltage variation will result in varying bias currents. Unless you also regulate the screens voltage off course, as screen voltages will largely determine the plate current.
rootz wrote: ↑Fri Feb 14, 2020 11:43 am
Diodes are reversed and centre tap is not needed; no ground connection there. Search for full wave voltage doubler. Here for example http://www.tubecad.com/january2000/page14.html You also won’t need high voltage caps, I think 2x 47u/160v will be enough for the doubler, 250v for the extra filtering. Keeps it small and cheaper
On the extra filtering part: seems like a good idea. You do not need -200V for full cutoff. I’d advise against just regulating the negative supply if you take the bias from it. That would make it resistant against line voltage variations. Screen and plate voltage won’t be though, so a line voltage variation will result in varying bias currents. Unless you also regulate the screens voltage off course, as screen voltages will largely determine the plate current.
Thank you very much for the correction relating to the diodes and the center tap. I have amended the schematic accordingly.
Also noted with thanks regarding not needing high voltage caps. I went with larger ones just because on one of those blurry pics from the amp the 3 caps seem to be fairly large (and also equal size). But good to know I can go lower voltage.
Regarding the extra filtering, I am not sure I follow exactly your advice, apologies. I did not add a voltage regulator, and I understand you advise against it as well. So I assume my layout is good?
You do not have the required permissions to view the files attached to this post.
That schematic looks good. The 2 caps in the doubler both see half the rectified voltage. That would be around 90VDC, so I think 160V rating would be plenty there. The cap most to the right in your schematic will see the full rectified voltage, especially under startup conditions. In that case there is no current being drawn from the circuit, so no voltage drop over the dropping resistor. Spec that one accordingly, probably 250V. Using 2 47uF caps in the doubler, a 4k7 dropper and 22u filter cap, would give you around 25mV of residue ripple (if my math is correct). Perfectly in phase when presented to the power tube grids, so I'm rather confident you won't hear any noise from the negative supply.
About regulating which I did on my Hifi amps and is much off topic here.
Quote from Tubelab (George, a much respected builder with tons of experience) at diyaudio. Please note there are two quotes in bold in this quote : Quote:
The "gotcha" is that O/P tube B+ must be regulated too.......Yes. Regulate both or regulate neither
In a TRIODE amplifier this is true since the plate CURRENT depends on the plate and control grid VOLTAGE.
In most PENTODE amplifiers the plate CURRENT depends on the plate and SCREEN grid VOLTAGE.
So in a triode based output stage, you would regulate the plate and grid bias supplies.....these are really the only two choices anyway.
In a pentode based output stage, you would regulate the control grid and screen grid supplies. Regulating the plate supply is optional.
Quote:
the result will be much the same either way, but one is much simpler than the other
In most cases this is true. If the two supplies are derived from the same transformer, or otherwise derived in a manner that they are both linearly dependent on the line voltage, then neither, or both should be regulated, from a tube current stability point of view.
In most cases the power supply is in the signal path, and thus should be a low impedance source for all frequencies that the amplifier will be reproducing. A well designed unregulated supply with good capacitors is usually quite low over all important frequencies.
Often some regulated supplies may present an unusual output impedance spike when presented with a dynamic load from an amp producing music at high volumes into a speaker. This can be caused by the feedback path in the regulator itself. A poorly designed, and tested for all expected operating conditions, regulator is sometimes worse than no regulator at all.
End of quote
To add to that: there are some highly experienced, very helpful and good out of the box thinkers on diyaudio. Tubelab being one of them, as are Sy, DF96 and many others. Many good reads there if you are interested.
Last edited by rootz on Fri Feb 14, 2020 3:09 pm, edited 3 times in total.
You also had a question about where to connect the g2 decoupling cap of a small signal pentode. In this particular case, you can do it either way: to the cathode or to ground. For most frequencies it won't matter, because of the bypass cap on the cathode. AC is at ground potential with a bypass cap, so the screen grid will be too (off course this logic fails with partially bypassed cathodes, 1k5//10uF counts as fully bypassed at guitar frequencies in my book).
From a practical point of view you might like decoupling g2 to the cathode much better then other options. You could essentially wire a 100nF/630V cap right on the tube socket. Doesn't get much simpler than that.
rootz wrote: ↑Fri Feb 14, 2020 2:51 pm
That schematic looks good. The 2 caps in the doubler both see half the rectified voltage. That would be around 90VDC, so I think 160V rating would be plenty there. The cap most to the right in your schematic will see the full rectified voltage, especially under startup conditions. In that case there is no current being drawn from the circuit, so no voltage drop over the dropping resistor. Spec that one accordingly, probably 250V. Using 2 47uF caps in the doubler, a 4k7 dropper and 22u filter cap, would give you around 25mV of residue ripple (if my math is correct). Perfectly in phase when presented to the power tube grids, so I'm rather confident you won't hear any noise from the negative supply.
Thanks you very much Rootz. I have amended the schematic with your value for future referece.
Also interesting off topic intel!
I appreciate the help on the pentode wiring as well. Good to know that I can go either way. As far as the practicality you are right, wiring the 0.1uF (630V) straight at the socket would be easy (If not using the scattered SSS style layout).
decoupling cap of a small signal pentode. In this particular case, you can do it either way: to the cathode or to ground.
You do not have the required permissions to view the files attached to this post.
Make sure you use filter caps on the low level side (47/350 stacked) so they can withstand the cold start voltages.
Thanks Rootz for the final touches on the C- circuit. I also derived the current calculation and the current of heaters from the original transformers will be fully used by the Power tubes. So the additional filament transformer will be used for all the pre amp tubes.
And based on the HAD pic attached one can see the 4 power supply caps. Since I am yet to see an SSS with a higher value than 100uF, that is what I went for. The same pic also shows 4 additional supply caps, two smaller and 2 larger. The smaller might be the Fuchs stacked ones? Also he did not go for a precision supply.
Now a question regarding the power amp and preamp supplies. In one of my amps the Screens were at 452.9V and the 6.8K dropping resistor after dropped the voltages for the PI to around 300V. I am wondering if 40K would maybe be enough to drop the 700V to around 400V which this PI needs, without the need for the ground reference like the Tan Music Man? Or would it be advisable to add the 820K to ground regardless.
You do not have the required permissions to view the files attached to this post.
That's not what Andy meant. He meant stacking preamp filter caps like in the precision power supply. That is where the cold start voltages will be high enough to kill 500V caps.
Better still, or more to my liking at least, would be to protect the lower side with some zeners, to maximise at 450V or thereabouts. That would keep 500V caps safe in any condition and have no impact on the PS when the tubes are warm, because they will drag the voltage below the zener threshold.
Also, there is a centre tap connected in your high voltage winding. That tap is not on the taro you're looking at and, most of all, not needed.
I'm going to draw you a power supply that should get you real close to what you want/need. 40k will be enough for the PI.
rootz wrote: ↑Fri Feb 14, 2020 5:58 pm
Also, there is a centre tap connected in your high voltage winding. That tap is not on the taro you're looking at and, most of all, not needed.
Not only is it not needed, it renders the circuit useless. The first thing that happens is the two diodes connected to ground will pop. And when the standby switch is closed all the e-caps in the lower half will slowly explode.
Thank you for the note on the central tap and thanks Rootz for the corrections on the stacked observation and for helping on drawing a correct power supply.
Based on this pic from overland express below and on my post above I think the entire supply is as follows:
4 capacitors Cap Can (4x 15 to 20uF): probably Triode, Pentode, Send/Return, Mixer
2 small capacitors, probably stacked like Andy suggested
2 normal size capacitors: PI, PI Driver
4 power supply capacitors
Missing is the reverb driver. I am thinking that maybe Like the mayer where the reverb driver is without a capacitor
Tremolo needs a negative and also 415V in plates (maybe same as PI)?
You do not have the required permissions to view the files attached to this post.
Not verified against one of the pictures you have. I leave that up to you. This is based on the info from the schematic Aaron posted in combination with what we know about the MM amps/Tony's info. And of course your info here.
Still thinking about why of if HAD didn't use precautions against over voltage on Start up/cold tube conditions. Multi cans normally share their cathodes and are mostly max 500V. Even with all the time constants (big ass dropper right before the PI) I'd think that the PI cap sees in excess of 600V without precautions. I use limiting zeners. They do dissipate close to 1W for a short period though. Another plus is that this won't put 600V on the plates of some cold preamp tubes.
You could take the negative for the term from the junction of the two caps in the doubler. Might need some extra filtering. Or you could take from the filtered negative supply and drop some more voltage.
Another thing that concerns me slightly: very high voltage on the grid of the 12DW7 CF at cold situations and with a cathode at ground potential. Nice recipe for arcing. At warm conditions the cathode is at +200V, violating the heater to cathode voltage in any case with a positive AC signal.
You do not have the required permissions to view the files attached to this post.
The Zeners are a good idea I used them in the Classic PS along with the 820K to set the time constant there. I've often wondered why Dumble didn't do this on his Classics without the stacked caps?. It's also a good idea to put some after the V1 filters to protect the FET from short spikes at start-up. Personally I would go 100uF on the plates and 20uF on the preamps. If you want to raise the screens a little 40uF for a little better regulation might be a good idea. I doubt you will be able to tell a difference between 2x6550 or 4 maybe at real high volumes. Even SRV couldn't make that amp break up playing as LOUD as he did.
BTW. I've seen some Dumbles where the preamp dropper was either a 35k or a 40k. You may want to get both while your ordering them. I started with the 35 and then had to raise it to a 40. Depends on where things settle out.
Tony
" The psychics on my bench is the same as Dumble'"
I doubt you could even hear the difference between 2x6550 or 4x6L6.
Good idea about FET protection. I did something similar in an LND150 effects loop; zenders to protect them against voltage spikes.
With the zener string I’ve drawn the big dropping resistors in front of it dissipate 14W during start up. I’d use two 10W resistors, but that might not be ‘cloning’.
When it comes to cloning: how much do we actually know for sure about the silver alligator? Is the proposed power supply actually accurate (apart from the zener string)? Did the amp really sport 4x6550 @ 700V using HD130 transformers?
Thanks Tony and Rootz. I am definately going for a 2x6550 but I am trying to document a 4x6550 (because of the original). Also trying to build this up based on the somewhat limited info available.The amp does not have FET, so one less thing to think about.
I am attaching a close up of one of those blurry shots from an old thread which shows the 4 supply caps. One can clearly see 2x 47uF (450V) and a resistor in between. The two smaller caps might be 22uF or 47uF in a lower rating? Maybe 350v or 250V? I added 47uF/350V to the schematic solely based on Andy's comment
Tony taking you suggestion into account, maybe he has the 2x47uF (450V) for screens and reverb diver as per my schematic? Maybe the resistor between the 2x47uF (450V) is the dropping to the reverb which would be somewhere between 470R and 2K?
As far as the cold side, would it make sense that he stacks the smaller caps on the PI (like my schematic) and then go into the cap can with a non precision, classic style supply? I would guess the cap can is rated at most 500V.
Unfortunately with this configuration we are one cap short for the PI Driver
I will probably add those zeners to my schematic. Is the 820K to ground required?
You do not have the required permissions to view the files attached to this post.
: