GM70 PP
- izzy wizzy
- Old Hand
- Posts: 1496
- Joined: Fri Nov 02, 2007 7:02 pm
- Location: Auckland NZ
- Contact:
#1 GM70 PP
The 813 amp has become a dead end as the 813 is now hard to get easily at reasonable prices. Folk round here tell me the 813 isn't that great incomparison to the GM70. I want to have enough valves to feed this thing in the coming years and GM70 are reasonably priced and easy to get. So here we are at converting the 813 PP amp into GM70 PP. Driver will remain 4P1L so this will be an all Russian DHP/DHT amplifier.
First I have to get all the parts. Waiting on SMD caps for Andrew's VCCS supplies to power the 4P1L and I can build the supplies. Probably will be tested in the current 813 amps.
GM70s and sockets have arrived. I'm going to power them from SMPS again but this time, 1 per valve. The Meanwell supplies (on their way) are variable and in the end, I hope to power the fils from a modified version of Andrew's VCCS much like Mark has done for the GK71. But for starters, I'll wind them down to about 22V and drop some volts with resistors and feed the fils via a 8mH CM choke. That's about all that's needed for the conversion.
The question I have is what is the purpose of a vitual centre tap with resistors across the GM70 fil to gound when it's being DC powered? I ask as in my head, the voltage gradient across the filament must favour the signal to one end. Or in the case of AC, the music, is that developed evenly across the filamant and so then I can see the point of the centre tap provided by the resistors. If any of that makes sense, I would welcome comments and insight.
Cheers,
Stephen
First I have to get all the parts. Waiting on SMD caps for Andrew's VCCS supplies to power the 4P1L and I can build the supplies. Probably will be tested in the current 813 amps.
GM70s and sockets have arrived. I'm going to power them from SMPS again but this time, 1 per valve. The Meanwell supplies (on their way) are variable and in the end, I hope to power the fils from a modified version of Andrew's VCCS much like Mark has done for the GK71. But for starters, I'll wind them down to about 22V and drop some volts with resistors and feed the fils via a 8mH CM choke. That's about all that's needed for the conversion.
The question I have is what is the purpose of a vitual centre tap with resistors across the GM70 fil to gound when it's being DC powered? I ask as in my head, the voltage gradient across the filament must favour the signal to one end. Or in the case of AC, the music, is that developed evenly across the filamant and so then I can see the point of the centre tap provided by the resistors. If any of that makes sense, I would welcome comments and insight.
Cheers,
Stephen
#2 Re: GM70 PP
I have argued that a centre tap with DC will help reduce any residual noise. But I think it practice its not worth doing and may be worst as you are adding extra resistors in the circuit.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- izzy wizzy
- Old Hand
- Posts: 1496
- Joined: Fri Nov 02, 2007 7:02 pm
- Location: Auckland NZ
- Contact:
#3 Re: GM70 PP
Thanks Nick. It seems to have become standard practice if I look around the interweb but then couldn't figure it out with DC fils. AC sure. I'm not in favour of adding more resistors for the sake of it and don't fancy spending the money for a comparison if there's no real need. But I question it when I see good technical people like Pete Millet and Andrea Ciuffoli etc doing it albiet with SE amplifiers if that makes any difference.
I'd like to undertand what residual noise you mean? And do I take it that you've tried the with and without? Or is that a hunch that it may be no better in practice? Especially with a decent supply like Andrew's VCCS?
Cheers,
Stephen
#4 Re: GM70 PP
Its based on a mathematical analysis of how the centre tap works I did some years ago (there is a copy of the maths somewhere on here). As to if there is any noise , don't know. the odd thing is, its hard to find a explanation how a hum pot works, so I worked out my own answer to that and its applicable to DC as well.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
- Paul Barker
- Social Sevices have been notified
- Posts: 8996
- Joined: Mon May 21, 2007 9:42 pm
#5 Re: GM70 PP
On valve choice, 813’s on sale sh which I bought were wildly different transconuctance , I expect due to used ones, because of their intended use got a real pasting. Nos barely available. To find four used matching near impossible.
Get the new gk71 or gm70
Get the new gk71 or gm70
"Two things are infinite, the universe and human stupidity, and I am not yet completely sure about the universe." – Albert Einstein
-
- Thermionic Monk Status
- Posts: 5649
- Joined: Thu May 24, 2007 11:22 am
- Location: People's Republic of South Yorkshire
#6 Re: GM70 PP
I've always thought, perhaps erroneously, that hum pots were used with AC filaments to find the "balance" point across the filament potential difference where the hum was minimum. I've never used a hum pot but have used a pair of balancing create a centre point approximation of the PD.
I'm not sure why a hum pot or pair of resistors would be used with DC fils. That said, there's c20V across the GM70 filament and I've never really got my head around what the effect is on the operating point of the GM70 is by attaching the cathode resistor to either the "+" or "-" end of the filament, assuming common cathode. Perhaps with such a large filament voltage there is some merit in a centre tap? I haven't bothered with one though.
I'm not sure why a hum pot or pair of resistors would be used with DC fils. That said, there's c20V across the GM70 filament and I've never really got my head around what the effect is on the operating point of the GM70 is by attaching the cathode resistor to either the "+" or "-" end of the filament, assuming common cathode. Perhaps with such a large filament voltage there is some merit in a centre tap? I haven't bothered with one though.
- izzy wizzy
- Old Hand
- Posts: 1496
- Joined: Fri Nov 02, 2007 7:02 pm
- Location: Auckland NZ
- Contact:
#7 Re: GM70 PP
I suspect my 813s were 2nd hand but from someone reputable IIRC back in the day and so they were pretty close however this is why I now have 8 new GM70 sitting beside my desk ready to go as I pretty much came to the same conclusionPaul Barker wrote: ↑Sat Jul 13, 2019 7:57 pm On valve choice, 813’s on sale sh which I bought were wildly different transconuctance , I expect due to used ones, because of their intended use got a real pasting. Nos barely available. To find four used matching near impossible.
Get the new gk71 or gm70
![Smile :)](./images/smilies/icon_smile.gif)
Cheers,
Stephen
#8 Re: GM70 PP
Yes, thats what they do, but what I never found was a explanation of how they do it.I've always thought, perhaps erroneously, that hum pots were used with AC filaments to find the "balance" point across the filament potential difference where the hum was minimum.
I think its very easy to imagine that more electrons leave the end of the cathode that is more +ve. But I don't believe thats the case. The electrons don’t travel from the filament to the anode, they are boiled off into the space charge around the filament by the temperature of the filament. They then are attracted to the anode by the potential difference. But the space charge is a cloud of electrons, I I don't think there can be a potential difference from one part of the space charge to another, if there was the electrons would move to the +ve end removing the difference.I'm not sure why a hum pot or pair of resistors would be used with DC fils. That said, there's c20V across the GM70 filament and I've never really got my head around what the effect is on the operating point of the GM70 is by attaching the cathode resistor to either the "+" or "-" end of the filament, assuming common cathode. Perhaps with such a large filament voltage there is some merit in a centre tap? I haven't bothered with one though.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
-
- Thermionic Monk Status
- Posts: 5649
- Joined: Thu May 24, 2007 11:22 am
- Location: People's Republic of South Yorkshire
#9 Re: GM70 PP
Well that would make a lot of sense. And probably explains why a centre tap doesn't make much difference.
There are reports though of people preferring the cathode resistor attached to the + rather than -, suggesting there is some difference, but I've never tried it.
There are reports though of people preferring the cathode resistor attached to the + rather than -, suggesting there is some difference, but I've never tried it.
#10 Re: GM70 PP
I don’t know, I could imagine one end of the filament may sound different to the other, not because of the filament voltage, but physically, they may be different. I don’t; know, so can't say. The cause of the effect may not be the one that’s expected, or it may be, reality can be tricky.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
-
- Thermionic Monk Status
- Posts: 5649
- Joined: Thu May 24, 2007 11:22 am
- Location: People's Republic of South Yorkshire
#11 Re: GM70 PP
Indeed. But that's why this hobby is so interesting.
- pre65
- Amstrad Tower of Power
- Posts: 21400
- Joined: Wed Aug 22, 2007 11:13 pm
- Location: North Essex/Suffolk border.
#12 Re: GM70 PP
Interesting.
Whilst looking into this I found this on a Rod Coleman filament supply details sheet.
"Comparative listening tests usually show a preference for the use of the FILAMENT +
side of the filament used as the return-path for the cathode-current. Please feel free to
experiment with other connexions (centre-tap, FILAMENT - ) as return paths, but note
that the bias is affected: the DHT is biased 5V hotter (more anode current) for
FILAMENT + (5V filaments). Please adjust the grid bias or autobias resistor value,
before testing."
So, how would this apply to establishing the cathode resistor value on GM70/GK71 with a 20v filament ?
Whilst looking into this I found this on a Rod Coleman filament supply details sheet.
"Comparative listening tests usually show a preference for the use of the FILAMENT +
side of the filament used as the return-path for the cathode-current. Please feel free to
experiment with other connexions (centre-tap, FILAMENT - ) as return paths, but note
that the bias is affected: the DHT is biased 5V hotter (more anode current) for
FILAMENT + (5V filaments). Please adjust the grid bias or autobias resistor value,
before testing."
So, how would this apply to establishing the cathode resistor value on GM70/GK71 with a 20v filament ?
The only thing necessary for the triumph of evil is for good men to do nothing.
Edmund Burke
G-Popz THE easy listening connoisseur. (Philip)
Edmund Burke
G-Popz THE easy listening connoisseur. (Philip)
#13 Re: GM70 PP
My own experience was that removing those virtual cathode resistors was a huge improvement to the sound. Really quite surprised me at the time, I was just experimenting really, but it was night and day for me. My amp uses Andrews boards and is all PP DHT (26 - 71A - 45).izzy wizzy wrote: ↑Sat Jul 13, 2019 4:37 pm
The question I have is what is the purpose of a vitual centre tap with resistors across the GM70 fil to gound when it's being DC powered? I ask as in my head, the voltage gradient across the filament must favour the signal to one end. Or in the case of AC, the music, is that developed evenly across the filamant and so then I can see the point of the centre tap provided by the resistors. If any of that makes sense, I would welcome comments and insight.
Cheers,
Stephen
If you do decide to install them, obviously choice of value is important because you create a parallel path for the heater current so with high current valves you need to be careful on dissipation. Nick helped me get my head around this at the time (thanks again) and I have a spreadsheet if it helps.
#15 Re: GM70 PP
Had another one of those "Are we looking in the wrong direction" moments. While most DC supplies may be floating, I don’t thing they are balanced. At least not from the perspective of the audio signal path. So the fact that causing the signal to exit from one or the other end may well make a difference. Not because of the voltage on the filament, but the effect of that part of the DC supply on the signal as it passes by.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.