Black Gate VK up to 350V:
https://www.ebay.com/sch/i.html?_fr...LH_TitleDesc=0&_osacat=0&_odkw=black+gate+wkz
Black Gate WKZ up to 500V:
https://www.ebay.com/sch/i.html?_fr...&LH_TitleDesc=0&_osacat=0&_odkw=black+gate+wk
Elna Cerafine 350...500V capacitors also good ones.
euro21: I meant in the flesh, and in my hands. The prices are so insane that they might as well not be available. At that price, they have to be the best, right?
I have Cerafine on hand in voltages for cathode bypass. I also have Silmic II for cathode bypass as well. When it matters, I’m going use a big chassis, or more than one chassis, and use film caps.
A quick update on the 30/6P15P-EV amp...
I added more capacitance to the B+ node that feeds the type 30 tubes. They now each have the original small film cap, and an additional 33 uF per side. I didn't notice any obvious changes, but it also didn't seem to hurt the performance either, so I'll leave them in. I still haven't found any formula or rule of thumb relating to ideal capacitor size for a given node. I will send jhstewart9 a message and ask him about it.
I also changed the cathode resistors on the 6P15P-EV tubes back to their higher value for the original lower dissipation, and it sounds better this way than at higher dissipation.
I am enjoying the amp, and listen to it daily while working on finishing a couple of Nelson Pass projects started long ago but left unfinished due to an irrational fear of tapping aluminum heat sinks. I got over it by successfully tapping the ACA heat sinks, and will do the Aleph J next.
Once done with the ACA and Aleph J, I am going to do another 30/6P15P-EV amp, but this time with battery bias and filaments for the 30 tubes. I would also like to build this next 30/6P15P-EV amp so that it can serve as the front end for a large triode or triode connected pentode amp in another chassis. Perhaps i can do this by having both a preamp output and an output transformer output for speakers on a switch.
I added more capacitance to the B+ node that feeds the type 30 tubes. They now each have the original small film cap, and an additional 33 uF per side. I didn't notice any obvious changes, but it also didn't seem to hurt the performance either, so I'll leave them in. I still haven't found any formula or rule of thumb relating to ideal capacitor size for a given node. I will send jhstewart9 a message and ask him about it.
I also changed the cathode resistors on the 6P15P-EV tubes back to their higher value for the original lower dissipation, and it sounds better this way than at higher dissipation.
I am enjoying the amp, and listen to it daily while working on finishing a couple of Nelson Pass projects started long ago but left unfinished due to an irrational fear of tapping aluminum heat sinks. I got over it by successfully tapping the ACA heat sinks, and will do the Aleph J next.
Once done with the ACA and Aleph J, I am going to do another 30/6P15P-EV amp, but this time with battery bias and filaments for the 30 tubes. I would also like to build this next 30/6P15P-EV amp so that it can serve as the front end for a large triode or triode connected pentode amp in another chassis. Perhaps i can do this by having both a preamp output and an output transformer output for speakers on a switch.
Thanks Bill. I’ll check it out.
Edit: It’s Loesch! I’m always interested in what he has to say.
Bill Brown: I started reading and realized that I’ve read the article previously. I read it again anyway. This article refers to the power tube (300B) and not the driver tube when discussing “final capacitor”.
This is my understanding of how he works out the power tube capacitance and the maximum permissible total resistance of the power supply for the 300B...
1. Thorsten takes the 300B’s anode resistance, 700 Ohms, and says the the entire power supply impedance should not exceed one tenth of that resistance. That means the supply should have 70 Ohms or less of resistance at frequencies above 4 HZ.
2. He describes how a time constant of 40 mS or less is calculated as an acceptable recharge time on the final capacitor.
3. He derives a capacitor value of 560 uF by using uF=kOhms/mS uF=.07/40=560
4. kOhm X uF = mS is used again to find the supplies maximum permitted resistance. mS/uF=kOhm 40/560=.071 kOhms.
5. 70 Ohms as a maximum resistance for the entire power supply is close to impossible without using a regulated supply, so he relaxes the lowest frequency of interest to 14 Hz from 4 Hz, which leaves a new maximum resistance of 266 Hz.
This is my understanding of how he works out the power tube capacitance and the maximum permissible total resistance of the power supply for the 300B...
1. Thorsten takes the 300B’s anode resistance, 700 Ohms, and says the the entire power supply impedance should not exceed one tenth of that resistance. That means the supply should have 70 Ohms or less of resistance at frequencies above 4 HZ.
2. He describes how a time constant of 40 mS or less is calculated as an acceptable recharge time on the final capacitor.
3. He derives a capacitor value of 560 uF by using uF=kOhms/mS uF=.07/40=560
4. kOhm X uF = mS is used again to find the supplies maximum permitted resistance. mS/uF=kOhm 40/560=.071 kOhms.
5. 70 Ohms as a maximum resistance for the entire power supply is close to impossible without using a regulated supply, so he relaxes the lowest frequency of interest to 14 Hz from 4 Hz, which leaves a new maximum resistance of 266 Hz.
This should be...
5. 70 Ohms as a maximum resistance for the entire power supply is close to impossible without using a regulated supply, so he relaxes the lowest frequency of interest to 14 Hz from 4 Hz, which leaves a new maximum resistance of 266 Ohms.
I am not sure how he works out the 266 Ohms as a result of the change from 4 to 14 Hz.
5. 70 Ohms as a maximum resistance for the entire power supply is close to impossible without using a regulated supply, so he relaxes the lowest frequency of interest to 14 Hz from 4 Hz, which leaves a new maximum resistance of 266 Ohms.
I am not sure how he works out the 266 Ohms as a result of the change from 4 to 14 Hz.
These maths are theoretical, because every PSU with tube rectifier (even my mercury vapour ones, which have only few ten Ohm "resistance") crosses that impedance border.
The 99.9% of the 300B amps not use this high last capacitor.
IMHO over 100uF the tone becomes "lazy", even if I use BG or Cerafine.
BTW the "classical" 5U4GB-8uF-10H(58R)-100uF breaks this 70R border at 30Hz (simulation).
The 99.9% of the 300B amps not use this high last capacitor.
IMHO over 100uF the tone becomes "lazy", even if I use BG or Cerafine.
BTW the "classical" 5U4GB-8uF-10H(58R)-100uF breaks this 70R border at 30Hz (simulation).
Which is a reasonable choice as 99% of music made with real instruments has zero content below 25Hz. 25-30Hz is also a typical limit for full power of the OPT, including high quality transformers. To get down to 20Hz one has to use BIG transformers for the job at hand...
The 1% of music left can contain sparse notes of a pipe organ and (the super-rare) Octobass. Both I think can play only 1 or 2 notes below 20Hz. But more importantly, respect to the amplifier recovery, I don't think is gonna be a continuous beat.
So it's really all about Heavy Metal and Electronic music. Do you really listen to this stuff with a low power SE amp? Is it really worth the trouble? In such case, I would use a regulator or hook up a sub with its own amp.....as most loudspeakers also have lots of troubles going so low. It's not just the SE amp. It's everything being unsuitable.
Last edited:
The bottom E on a 4 string bass is around 40hz. 5 string basses have a low B at 31hz. Unless you're an organ/classical piano fan there's nothing of value going on below 31hz. Lowest note on a concert grand is 27.5hz. In any case, a lot of the sound of these low instruments is contained in the overtones at higher frequencies.
.
.
Last edited:
"This is my understanding of how he works out the power tube capacitance and the maximum permissible total resistance of the power supply for the 300B..."
Yes, I agree with the understanding leading to your listed points. I liked that the requirements he outlined were from "first principles." Though, as has been noted, the total PS impedance would be almost impossible to achieve without regulation, so compromises are necessary.
My thinking is that the calculations used to determine the (theoretical) size of the final output tube could be applied to preceding stages as well.
Bill
Yes, I agree with the understanding leading to your listed points. I liked that the requirements he outlined were from "first principles." Though, as has been noted, the total PS impedance would be almost impossible to achieve without regulation, so compromises are necessary.
My thinking is that the calculations used to determine the (theoretical) size of the final output tube could be applied to preceding stages as well.
Bill
Bill: I accidentally switched the formula over in my post above, I put in uF=kOhms/mS instead of the correct uF=mS/kOhms. I don't see why the formula used wouldn't be applicable to the pre stage tubes.
Using Loesch's method for 30 tubes...
The 30 tube has a plate impedance of 11000 Ohms. One tenth of that is 1100 Ohms.
Using the same time constant of 40 mS or less (based upon 4 Hz) as an acceptable recharge time on the final capacitor, uF=mS/kOhm uF=40/1.1=36
36uF is based on his 4 Hz criteria. I don't know how he works out the time constant for the derated 14 Hz, but in his case using the 300B, this goes down to 266 uF from 560 uF. Reducing the uF proportionately for the 30 results ib 17.1 uF for the derated 14 Hz.
What would be the result if the process is applied to the 30 Hz or so suggested by andyjevans and 45?
In my amps ultimate role as a mid/treble amp with a lower frequency of around 600 Hz, even my original miniscule 2.2 uF is likely far more than required. That said, any guidance on how he derives the time constant based on frequency would be much appreciated, as it is not explained in the article.
Edit: I think I found a calculator and explanation on the derivation based on frequency here...
RC Circuit Calculator
Last edited:
Putting Loesch's numbers for the 300B of 70 Ohms (one tenth plate impedance) and 4 Hz into the calculator results in a capacitance of 560 uF. Using 14 Hz results in 160 uF.
For the 30 tube, using 1100 Ohms (one tenth of its 11000 Ohm plate impedance) and 4 Hz results in 36 uF.
14 Hz results in 10 uF.
30 Hz results in 4.8 uF.
600 Hz results in .24 uF
Using the originally installed 2.2 uF capacitors on the 30 results in 65.8 Hz. This is too high for a full range amp, so the concerns of jhstewart9, andyjevans and FlaCharlie were well founded based upon full range output.
This amp now has a 33 uF capacitor added per channel in parallel to the original 2.2 uF capacitor per channel for a total of 35.2 uF. Given that I require 5 uF for 30 Hz, am I not better off removing the parallel electrolytic and film 35.2 uf and replacing it with a 5 uF film cap? This would satisfy the requirements if the amp is run full range (30 Hz and up), and it would greatly exceed the requirements if the amp is run at 600 Hz and up.
For the 30 tube, using 1100 Ohms (one tenth of its 11000 Ohm plate impedance) and 4 Hz results in 36 uF.
14 Hz results in 10 uF.
30 Hz results in 4.8 uF.
600 Hz results in .24 uF
Using the originally installed 2.2 uF capacitors on the 30 results in 65.8 Hz. This is too high for a full range amp, so the concerns of jhstewart9, andyjevans and FlaCharlie were well founded based upon full range output.
This amp now has a 33 uF capacitor added per channel in parallel to the original 2.2 uF capacitor per channel for a total of 35.2 uF. Given that I require 5 uF for 30 Hz, am I not better off removing the parallel electrolytic and film 35.2 uf and replacing it with a 5 uF film cap? This would satisfy the requirements if the amp is run full range (30 Hz and up), and it would greatly exceed the requirements if the amp is run at 600 Hz and up.
Last edited:
euro21: Using this calculator...
RC Circuit Calculator
100 uF on the node for a 300B results in a frequency of 22.7 Hz. This seems reasonable to me, if still not satisfying Loesch's 4 or 14 Hz criteria.
The above works out how to establish an appropriate node capacitance based upon frequency cutoff. Loesch then goes on to use 40 mS and the capacitance of the tube’s PS node to work out what the DCR of the entire power supply (I am assuming up to that point) should be kept under. In his 300B example, 40mS/560uF = .071 kOhms 70 Ohms is close to impossible.
40mS is strict criteria based on demanding genres. 4 Hertz is also a strict criteria. Using 25 Hz for the 300B results in 91 uF being required, and a power supply DCR of 440 Ohms or under. Using 100 ms and 91 uF results in a maximum power supply DCR of 1100 Ohms.
He ends the article with his tube regulated supply for his Legacy amp. It is tuned to have a DCR of 15 Ohms, but can be set up for less than 1 Ohm.
40mS is strict criteria based on demanding genres. 4 Hertz is also a strict criteria. Using 25 Hz for the 300B results in 91 uF being required, and a power supply DCR of 440 Ohms or under. Using 100 ms and 91 uF results in a maximum power supply DCR of 1100 Ohms.
He ends the article with his tube regulated supply for his Legacy amp. It is tuned to have a DCR of 15 Ohms, but can be set up for less than 1 Ohm.
There are typo in the cited PDF: mS, which is milli Siemens (conductance, S: A/V).
"Given the relative behaviors (I'm guessing now) I'd suggest that the DC Impedance should be selected such that with everything considered we get no worse than 40 milliseconds (or 0.04 Sec) as Time-Constant for recharging the PSU Cap. Now the Time Constant for milliseconds can be calculated from the following formula: kOhm X uF = mS"
The correct form is "ms" (millisecond).
"Given the relative behaviors (I'm guessing now) I'd suggest that the DC Impedance should be selected such that with everything considered we get no worse than 40 milliseconds (or 0.04 Sec) as Time-Constant for recharging the PSU Cap. Now the Time Constant for milliseconds can be calculated from the following formula: kOhm X uF = mS"
The correct form is "ms" (millisecond).
I think you need to test ..
Get signal generator (can be as simple as 'test tone file')
and either CRO or appropriate meter.
Without that, you will be 'blind'. Unless of course you are theorizing.. I see a lot of that.
Can I share my recent 12B4/LM431 shunt regulator where I nulled noise with a mass of metal attached to the grid ?
Get signal generator (can be as simple as 'test tone file')
and either CRO or appropriate meter.
Without that, you will be 'blind'. Unless of course you are theorizing.. I see a lot of that.
Can I share my recent 12B4/LM431 shunt regulator where I nulled noise with a mass of metal attached to the grid ?
Last edited: