Hi,
i was looking for a good GM70 SE project, and stumbled upon this blog:
GM-70 SE amplifier | Firebottles!
All the credits go to the Blog owner, named Vincet Duparc (still wonder whether he's on the forum?).
I tried to comment on his blog, but no answer came, perhaps these pages are no more curated.
Well, the PSU seems very well designed, regulated but no stabilized, but there are some components that are quite uncanny to me (what they are, how are they controlled.).
In the second attached schematic, the PSU, I circled in red my doubts:
i) I see two relays on the HT primary, one "tempo" shoud be a time delay one, that will be shorting out R27; same should apply to the inrush current limiter that shorts out the thermistor NTC2. Am I right? I think they should be 250VAC/3A, right?
ii) what are these F7/F8/F9 points before each mosfet ?
iii) in the second attached picture, there is another inrush relay that will short out NTC1, and ok, same as before.
In his blog pages, third schematic, there's an arduino controller for temperature and gm70 bias current (third attached picture). This drives a cascaded darlington ULN2003, that controls three relays, one for power, two for inrush limiting.
iv) are these three relays the same in the PSU schematic? I wonder why, if the controller switches them open, there still will be thermistors through which current flows... What am I missing?
If this is not the case (these 3 relays are not the ones in the PSU), where are these relays attached to?
I mean, the arduino schematic rationale seems to me as being 10V in, inputs from thermo sensors, inputs from bias current, and output what?
I really hope you can help me open my eyes, safety is always not enough with handling 1,1kV....
a big thank you!
Mike.
i was looking for a good GM70 SE project, and stumbled upon this blog:
GM-70 SE amplifier | Firebottles!
All the credits go to the Blog owner, named Vincet Duparc (still wonder whether he's on the forum?).
I tried to comment on his blog, but no answer came, perhaps these pages are no more curated.
Well, the PSU seems very well designed, regulated but no stabilized, but there are some components that are quite uncanny to me (what they are, how are they controlled.).
In the second attached schematic, the PSU, I circled in red my doubts:
i) I see two relays on the HT primary, one "tempo" shoud be a time delay one, that will be shorting out R27; same should apply to the inrush current limiter that shorts out the thermistor NTC2. Am I right? I think they should be 250VAC/3A, right?
ii) what are these F7/F8/F9 points before each mosfet ?
iii) in the second attached picture, there is another inrush relay that will short out NTC1, and ok, same as before.
In his blog pages, third schematic, there's an arduino controller for temperature and gm70 bias current (third attached picture). This drives a cascaded darlington ULN2003, that controls three relays, one for power, two for inrush limiting.
iv) are these three relays the same in the PSU schematic? I wonder why, if the controller switches them open, there still will be thermistors through which current flows... What am I missing?
If this is not the case (these 3 relays are not the ones in the PSU), where are these relays attached to?
I mean, the arduino schematic rationale seems to me as being 10V in, inputs from thermo sensors, inputs from bias current, and output what?
I really hope you can help me open my eyes, safety is always not enough with handling 1,1kV....
a big thank you!
Mike.
> ii) what are these F7/F8/F9 points before each mosfet ?
I suspect they are ferrite beads.
They can be helpful, but if the drain is properly bypassed, they will not be needed.
Good bypassing means: electrolytics as shown, plus stacked-film 100-220nF DIRECTLY at the drain, and with low-inductance loop all the way around to PS ground.
I suspect they are ferrite beads.
They can be helpful, but if the drain is properly bypassed, they will not be needed.
Good bypassing means: electrolytics as shown, plus stacked-film 100-220nF DIRECTLY at the drain, and with low-inductance loop all the way around to PS ground.
ok, thanks, I suspected that. First doubt solved. By the way, are you the same Rod Coleman of the regulated DHT module (also used in these schematic)?
A pleasure to meet you!
> i) I see two relays on the HT primary, one "tempo" shoud be a time delay one, that will be shorting out R27; same should apply to the inrush current limiter that shorts out the thermistor NTC2. Am I right? I think they should be 250VAC/3A, right?
Tempo is the start-up delay, to allow the filaments to be warmed up before the anode supply is transitioned to ON. I imagine that the primary NTC is then shorted at a later time, to allow full 1200V on to the anodes.
The filament supply is activated by SW4, which appears to be a mechanical switch - NOT controlled by the Arduino. That's OK.
It looks like a straightforward and robust control system - and it is certainly a good idea to have some protection for fixed bias amps.
The amplifier benefits from an excellent choice of Filament Regulators.
Tempo is the start-up delay, to allow the filaments to be warmed up before the anode supply is transitioned to ON. I imagine that the primary NTC is then shorted at a later time, to allow full 1200V on to the anodes.
The filament supply is activated by SW4, which appears to be a mechanical switch - NOT controlled by the Arduino. That's OK.
It looks like a straightforward and robust control system - and it is certainly a good idea to have some protection for fixed bias amps.
The amplifier benefits from an excellent choice of Filament Regulators.
So there's a separate on/off switch just for filaments? by looking at the enclosure this doesn't seem to be the case?
I think that sw4 is again a current inrush limiter for the filament transformer primary, that then bypasses NTC1... perhaps i'm wrong...
what is going to be a significant closing time for tempo? 3 seconds? 10 seconds?
so, what are the possible switches/relais that arduino controls? I looked at the PSU inside but no hints... they have to be circuit breakes though
yes, this was the main attraction to me
The amplifier benefits from an excellent choice of Filament Regulators.
oh, really sir?
Last edited:
When my Filament Regulator is used, there is no need for inrush limiting - the regulators are current-drive design, and there is no over-current at startup. This is beneficial to the filament, since rapid application of power can cause 10A peak current (severe re-crystallisation) and NTCs will not control it as accurately as the regulator.
So it may be that SW4 is just a configuration switch. You can forget it, if you use the regulators.
So it may be that SW4 is just a configuration switch. You can forget it, if you use the regulators.
This looks like a very well thought out schematic. The only thing i would change is the PN for the high voltage capacitor multiplier mosfets adding a transistor to pull away the gate charge during overcurrent conditions there only results in blown mosfets.
Perhaps add a Tube regulator there? you can accomplish the same thing that mosfet does with a tube just less performance but you can make up for that with more inductors in the power supply.
A time delay you can build out of a board i have in my free gerbers thread. No need to run an arduino in an amplifier. A 555 will provide a fine turn-on delay.
The slow start circuit on the +6.3V rail seems a bit overkill to me, but no harm no foul.
Rod's product is excellent i must say, all trough i offer my own regulator for transmitter tubes i think it doesn't give the same value for money as the Coleman regs, as its a quite expensive design to build.
Perhaps add a Tube regulator there? you can accomplish the same thing that mosfet does with a tube just less performance but you can make up for that with more inductors in the power supply.
A time delay you can build out of a board i have in my free gerbers thread. No need to run an arduino in an amplifier. A 555 will provide a fine turn-on delay.
The slow start circuit on the +6.3V rail seems a bit overkill to me, but no harm no foul.
Rod's product is excellent i must say, all trough i offer my own regulator for transmitter tubes i think it doesn't give the same value for money as the Coleman regs, as its a quite expensive design to build.
Yes, but controls the output of soft start/ tempo+2 inrush current limiters on transformers' primaries?
And how about the "power" relay arduino controlled? Where is it cabled to?
By the way, would you like to share your arduino program for your gm70? Or the scheme you used? Thank you!
An externally hosted image should be here but it was not working when we last tested it.
I'll need to dig out my old laptop for those files...I learned the program language with the help of another member on this site. You can probably do a search on my user name and find most the files. The Hammond IT has since been replaced by a Monolith Magnetics IT.
My GM 70 arduino has 4 relays, #1-Tube filaments, #2-Bias supply/driver supply, #3-GM soft start, #4- Gm70-Run (1 second after soft start). I use several of the inputs/outputs for monitoring, LED status lights, reset, start so it got a little complicated at times...
I built a voltage bridge to monitor the cathode bias and it looks for a range before going into protection mode. There's several other ways to accomplish the same tasks.
The gentleman that helped me now sells the boards and will probably help you write the program. I used a prototype board and found some relay boards off Ebay.
How to build a 21st century protection board check post 1502
AMP CONTROL BOARDS – Virtual Zero Audio Store
Last edited:
Nice amps everyone!
I've got over 12,000 hours of service out of my pair of GM70 amplifiers over the past 9 years without a single GM70 failure. (About 11 pairs so far and the current pair thanks to 9 months of working at home now have at least 1600 hours on them and will be replaced in December sometime.)
These amps have Rod's filament regulators - which I can highly recommend as both effective, reliable and minimal sonic footprint.
Other than making sure the bias comes up really fast I have taken no precautions, the bias comes up, the plate voltage shortly thereafter and finally the filaments.
The filaments are pure tungsten so not subject to stripping of an oxide layer, they're very poor emitters which is why they run 60W of filament power.
They are very linear however, and I think they are a very rational choice over current mediocre 845 production. (They are also extremely reliable as you would expect from Soviet military surplus.)
At $25 - $30 a tube I'm not willing to invest in a lot of protection circuitry and the one mistake I made ages ago the fuse on the plate transformer primary took care of that without damage to that particular amp. (There are a small number of other units around built by friends or built by me for people who don't build - they've been reliable.)
I estimate 5 hours a day average operation except on week ends where it might be 8 - 10 hours, and holidays where it is often more like 12 hours.
Ulyanov GM70 seem to last a bit longer than the ones made by Reflektor.
I went through my copper phase (3 sets) but these amps now do the bass in a 3 way active system so I use graphite plate.
I've got over 12,000 hours of service out of my pair of GM70 amplifiers over the past 9 years without a single GM70 failure. (About 11 pairs so far and the current pair thanks to 9 months of working at home now have at least 1600 hours on them and will be replaced in December sometime.)
These amps have Rod's filament regulators - which I can highly recommend as both effective, reliable and minimal sonic footprint.
Other than making sure the bias comes up really fast I have taken no precautions, the bias comes up, the plate voltage shortly thereafter and finally the filaments.
The filaments are pure tungsten so not subject to stripping of an oxide layer, they're very poor emitters which is why they run 60W of filament power.
They are very linear however, and I think they are a very rational choice over current mediocre 845 production. (They are also extremely reliable as you would expect from Soviet military surplus.)
At $25 - $30 a tube I'm not willing to invest in a lot of protection circuitry and the one mistake I made ages ago the fuse on the plate transformer primary took care of that without damage to that particular amp. (There are a small number of other units around built by friends or built by me for people who don't build - they've been reliable.)
I estimate 5 hours a day average operation except on week ends where it might be 8 - 10 hours, and holidays where it is often more like 12 hours.
Ulyanov GM70 seem to last a bit longer than the ones made by Reflektor.
I went through my copper phase (3 sets) but these amps now do the bass in a 3 way active system so I use graphite plate.
Why is the E186F Suppressor grid not connected to any DC return?
If the performance changes from cold to warm; or when it is at low signal volume versus mid or high volume, or when the humidity is 95%, or from one day to the next . . .
The suppressor grid, and the cap to ground let it be anything it wants to be.
Float, Float, Float. It may rise or fall to the occasion.
If the performance changes from cold to warm; or when it is at low signal volume versus mid or high volume, or when the humidity is 95%, or from one day to the next . . .
The suppressor grid, and the cap to ground let it be anything it wants to be.
Float, Float, Float. It may rise or fall to the occasion.
Last edited:
Suppressor grid is grounded, i see.
What's the point in leaving if floating with a capacitor soldered in?
Super duper! I'll have a look at your files, trying to understand some, it's a new language for me... At the moment I'm just understanding this PSU... Thanks anyway
Thank for your answer... Noticed sonic differences between copper (€150 each) and graphite (30€ each) plates?
I know these tubes are very stable, but prefer avoiding problems with 1kV floating around, for my wife, my newborn daughter and the two small cats here...
Michelag,
I am coincidentally looking at this same design/schematic to begin a longer term build. First research and understanding - breaking down the schematic etc, then focus on the PSU (similarly as a separate chassis) and eventually move into the amp.
Quick question - are you considering building the PSU into boards as the original designer did - I am intending to do something along those lines not only to better layout but also spend additional time on planning on checking. Do you know what sort of process would be adequate to implement to make these boards considering the voltages at B+ from power transformer?
I am coincidentally looking at this same design/schematic to begin a longer term build. First research and understanding - breaking down the schematic etc, then focus on the PSU (similarly as a separate chassis) and eventually move into the amp.
Quick question - are you considering building the PSU into boards as the original designer did - I am intending to do something along those lines not only to better layout but also spend additional time on planning on checking. Do you know what sort of process would be adequate to implement to make these boards considering the voltages at B+ from power transformer?