I have yet another newcomer's question about construction practices. I'm in the process of assembling a Pete Millett "Engineer's Amp", and wonder if I can successfully stack coupling caps in series to raise the total voltage rating.
I discovered a local surplus store which sells military and industrial metallized film caps for pennies on the dollar. I've been squirreling away bags of Wima and Epcos MKP10s and MKTs, as well as Cornell-Dublier metallized polycarbonate caps. I've found that these all work very well as signal caps. However, most are rated for 400VDC, and now I need some with 600 volt (or higher) ratings.
I'd like to use two .47uF, 400v caps in series to make a .235uF, 800v assembly. If I put these together with some 250K equalizing resistors, will they work properly as interstage coupling caps? Or will the multiple-component assembly degrade the signal, and introduce unwanted distortions?
If you've tried this before, I like to hear about your experience and results. Thanks!
I discovered a local surplus store which sells military and industrial metallized film caps for pennies on the dollar. I've been squirreling away bags of Wima and Epcos MKP10s and MKTs, as well as Cornell-Dublier metallized polycarbonate caps. I've found that these all work very well as signal caps. However, most are rated for 400VDC, and now I need some with 600 volt (or higher) ratings.
I'd like to use two .47uF, 400v caps in series to make a .235uF, 800v assembly. If I put these together with some 250K equalizing resistors, will they work properly as interstage coupling caps? Or will the multiple-component assembly degrade the signal, and introduce unwanted distortions?
If you've tried this before, I like to hear about your experience and results. Thanks!
Yes, you can, however I'd be careful to use large value shunt resistors 10Meg or greater so as not to effect the bias voltage of the grid in the driven tube as much.
For instance if you have a driving stage anode voltage of 200V and a grid voltage at gnd with a 470K grid resistor, and you use the two 250K resistors across your two caps, you will try to force the grid to near 100V!
Two 10M resistors will result in a 40:1 ratio so you will upset the grid by .47/(10*20+.47)=4.592V. Still quite a shift in operating point. You may have to change your biasing to compensate for this.
Two 22 Meg resistors will only offset you by .47/(44+.47) * 200 = 2.114V. Still significant.
For instance if you have a driving stage anode voltage of 200V and a grid voltage at gnd with a 470K grid resistor, and you use the two 250K resistors across your two caps, you will try to force the grid to near 100V!
Two 10M resistors will result in a 40:1 ratio so you will upset the grid by .47/(10*20+.47)=4.592V. Still quite a shift in operating point. You may have to change your biasing to compensate for this.
Two 22 Meg resistors will only offset you by .47/(44+.47) * 200 = 2.114V. Still significant.
I would thinks so, but I would expect that to be more complex than correcting the bias of the driven tube.
What kind of bias is currently used on the driven tube?
Could you post a schematic?
If straight cathode bias, it is just a matter of changing the cathode resistor value to correct the bias point.
If fixed bias, a change in bias voltage would be needed.
What kind of bias is currently used on the driven tube?
Could you post a schematic?
If straight cathode bias, it is just a matter of changing the cathode resistor value to correct the bias point.
If fixed bias, a change in bias voltage would be needed.
OK, Pete has a 100K to the -25 to -30V grid bias supply. If you use 10M resistors in series you have ~130V to 20M to the grid, and 100K to -25V.
This gives you an offset of (.1/(20+.1))*(130 - (-30)) = 0.79V
There is a good chance the bias adjustment will be able to take care of this offset. If it does not and you can't get to -25V at the grids, change R28 and R40 to 9K1 ohms and it should.
You need the 10M resistors with at least a 200V working voltage ratings (I'd go higher) as at power up you will have 340V+60V (B+ plus Bias) available to them. Since they are in series they will share this voltage.
Another reason for using such high resistor values is to prevent the change from effecting the loading of driver stage. The resistors are in parallel with the plate for ac analysis.
This gives you an offset of (.1/(20+.1))*(130 - (-30)) = 0.79V
There is a good chance the bias adjustment will be able to take care of this offset. If it does not and you can't get to -25V at the grids, change R28 and R40 to 9K1 ohms and it should.
You need the 10M resistors with at least a 200V working voltage ratings (I'd go higher) as at power up you will have 340V+60V (B+ plus Bias) available to them. Since they are in series they will share this voltage.
Another reason for using such high resistor values is to prevent the change from effecting the loading of driver stage. The resistors are in parallel with the plate for ac analysis.
Last edited:
Seems messy to me, and those 10M resistors add a degree of inelegance and compromise that I would find unacceptable as well as making the bias voltage a function of both the bias supply and variations in the plate voltage of the preceding stage. You've made the very definition of the leaky coupling cap on the verge of needing to be replaced - and deliberately! 
Generic 600V coupling caps aren't that expensive - I'd go looking for some.
Just my
Generic 600V coupling caps aren't that expensive - I'd go looking for some.Just my
Actually they are in series with the grid resistor and in parallel with the coupling cap XC, the whole combination including the miller capacitance of the driven stage then is in parallel with the plate load resistor of the driver stage - normally they would be inconsequential except at very low frequencies where they might introduce a step in the frequency response of the network.
IMHO still not a good idea unless gain to DC is required.
Kevin, I think I'll go with your "simple-and-straightforward-is-better" approach.
It was tempting to pair up some 15-cent metallized polycarbonate caps, but there are too many variables to make it worth the time expended.
Instead, I'll be using some Tobe-Deutschmann film-in-oil caps rated at .1uf/1000v. These are easily-overlooked military surplus metal cans which have amazingly low ESR and parasitic inductance (<.1nH). They look like miniature motor run caps, and were originally used for filtering on mobile generator trucks.
If you can find some at a surplus vendor, they can usually be had for about a buck apiece, and sound better than almost anything I've heard at 30 times the price.
It was tempting to pair up some 15-cent metallized polycarbonate caps, but there are too many variables to make it worth the time expended.
Instead, I'll be using some Tobe-Deutschmann film-in-oil caps rated at .1uf/1000v. These are easily-overlooked military surplus metal cans which have amazingly low ESR and parasitic inductance (<.1nH). They look like miniature motor run caps, and were originally used for filtering on mobile generator trucks.
If you can find some at a surplus vendor, they can usually be had for about a buck apiece, and sound better than almost anything I've heard at 30 times the price.
motor run caps
I'm using plastic film motor run caps for my tube b+ filtration. 50uf 440 VAC is about $15, and when I got them they were rated 1000 VDC. Metal cans, spade lugs on top. Not electrolytic, so don't have to match the voltage within 25% over. Will never dry out and need to be replaced. Pretty huge, though. I'm using Genelec brand ones made in Mexico.
I'm using plastic film motor run caps for my tube b+ filtration. 50uf 440 VAC is about $15, and when I got them they were rated 1000 VDC. Metal cans, spade lugs on top. Not electrolytic, so don't have to match the voltage within 25% over. Will never dry out and need to be replaced. Pretty huge, though. I'm using Genelec brand ones made in Mexico.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.