I decided to make GM70 SE amp several years ago. I have been reading about such projects, drawing schematics, collecting tubes, chokes, caps, cables and looking for design inspiritation since then Every end of the summer I said to myself: "Now is the time to start", but I quit then..
Last week I spoke to my colleague and he asked if I could make an amp for him. "Yes, sure, why not. Winter is coming, rains start pouring down so I will have more time to spend".
Am I afraid? A bit. I feel respect for HV because 500V is the highest voltage used in an amp so far.
Project specification (assumptions, limitations, requirements will be written next).
Last week I spoke to my colleague and he asked if I could make an amp for him. "Yes, sure, why not. Winter is coming, rains start pouring down so I will have more time to spend".
Am I afraid? A bit. I feel respect for HV because 500V is the highest voltage used in an amp so far.
Project specification (assumptions, limitations, requirements will be written next).
When I was building my GM70 amps I went out on a limb and tried a number of things like IT drive and a very odd single stage driver built around a triode connected D3A. The two amplifier chassis sat 90% complete for a year while I waffled about whether I should pour more money into this ill-advised project.
Fear of high voltages, and the potential disappointment of a design that failed to deliver were part of the issue.
I had originally planned on complex tube based regulated supplies and then decided to go with a very simple solid state bridge rectified CLCLC based HV supply. Bias and filament supplies followed. I completed the supplies in a matter of a week or so once I decided to proceed. (Had to wait for some custom power transformers from Edcor.)
When I finished the build out, did the preliminary checks, and fired the amps up for the first time I could have kicked myself in the head. They were clearly better than anything I had listened to up to that point, and a few minor tweaks were all that were required to address some minor issues.
They are now the only amps in the listening room (rest sold) and I certainly have no regrets having built them, and they have proved extremely reliable and unfinicky as well.
Get going guys!
Fear of high voltages, and the potential disappointment of a design that failed to deliver were part of the issue.
I had originally planned on complex tube based regulated supplies and then decided to go with a very simple solid state bridge rectified CLCLC based HV supply. Bias and filament supplies followed. I completed the supplies in a matter of a week or so once I decided to proceed. (Had to wait for some custom power transformers from Edcor.)
When I finished the build out, did the preliminary checks, and fired the amps up for the first time I could have kicked myself in the head. They were clearly better than anything I had listened to up to that point, and a few minor tweaks were all that were required to address some minor issues.
They are now the only amps in the listening room (rest sold) and I certainly have no regrets having built them, and they have proved extremely reliable and unfinicky as well.
Get going guys!
Thanks Kevin!
I will get to it after I finish another couple of projects that I have started. I realize now that 845 tubes are out of my budget and that the GM70 will work perfectly well. Below is the chassis and power supply I will use with the Magnequest FS 100 output transformers. Although I will have to rewire it and update the PS caps, the Ballantyne engineers have done much of the critical power supply design and safety engineering already. A Front Panel Express plate (or plates) will cover the original top.
I will get to it after I finish another couple of projects that I have started. I realize now that 845 tubes are out of my budget and that the GM70 will work perfectly well. Below is the chassis and power supply I will use with the Magnequest FS 100 output transformers. Although I will have to rewire it and update the PS caps, the Ballantyne engineers have done much of the critical power supply design and safety engineering already. A Front Panel Express plate (or plates) will cover the original top.
Attachments
The GM70 may well prove to be more reliable, I have no idea what current 845 production quality is like. So far I have only succeeded in wearing them out, no catastrophic behavior.
I will say I weed out the ones with bubbles in the glass. I may just be being ridiculous, but I do not know anything about glass and its integrity considerations. I figure 125W of plate dissipation and 60W filaments means a lot of stress. At one point I measured the glass temp and as I recall it was a pretty balmy 175C. (!) So I figure a breach through a bubble is possible and catastrophic. I've seen at least two tubes out of the 16 or so I have purchased so far with this issue. I have both copper plates and graphites, and lately I think I have come to prefer the graphite version overall. Never seen a bubble in the envelope of a copper plate GM70 but I only have 3 pairs.
I will say I weed out the ones with bubbles in the glass. I may just be being ridiculous, but I do not know anything about glass and its integrity considerations. I figure 125W of plate dissipation and 60W filaments means a lot of stress. At one point I measured the glass temp and as I recall it was a pretty balmy 175C. (!) So I figure a breach through a bubble is possible and catastrophic. I've seen at least two tubes out of the 16 or so I have purchased so far with this issue. I have both copper plates and graphites, and lately I think I have come to prefer the graphite version overall. Never seen a bubble in the envelope of a copper plate GM70 but I only have 3 pairs.
Project specs (1st approach):
1. Technical
* single ended parallel feed output stage GM70
* 2 stages
* Si devices allowed e.g. as CCS's or buffers
* should drive a pair of ELACs FS 207.2 JET in 70 m2 room
* monoblocs with a remote volume knob
2. Visual
* clean - just a flat surface and one GM70 tube, caps & xformers hidden
* light and thin
* modern material used, e.g. Ceramic glass
* minimal - look at Stixx's designs which inspire me
3. Functional
* 3 sources: CD, tuner and a flac player
* built-in display (vol/source) lit while operational, finger touch or proximity activated
* standby timer
1. Technical
* single ended parallel feed output stage GM70
* 2 stages
* Si devices allowed e.g. as CCS's or buffers
* should drive a pair of ELACs FS 207.2 JET in 70 m2 room
* monoblocs with a remote volume knob
2. Visual
* clean - just a flat surface and one GM70 tube, caps & xformers hidden
* light and thin
* modern material used, e.g. Ceramic glass
* minimal - look at Stixx's designs which inspire me
3. Functional
* 3 sources: CD, tuner and a flac player
* built-in display (vol/source) lit while operational, finger touch or proximity activated
* standby timer
I've built SET's off of 2A3, GU-50 (triode-strapped), 6C33C, and GM-70 (among others), and I'll tell you this: GM-70 have been among the most reliable tubes I've ever used. I was originally going to make a dual-socketed GM-70 amp capable of running 845's as well, but after playing with the Shuguang 845's, I decided not to bother. Not that the 845's were bad, but the GM-70's trounced 'em in terms of linearity, sturdiness, and cost. Can't say enough about 'em.
Don't worry. I've had a couple 6C33C's that I was suspicious of, but they held up (and are holding up) just fine...and GM-70's run cool compared to 6C33C's — (God, those things are hot).
Overall I am very happy with my decision to build mono blocks based on the GM70, it has proven quite reliable, and provides excellent performance and consistency. My second set of tubes are not far now from the 1K hour mark and have not been a cause for concern.
My concern about envelope temperature only relates to the few I have with inclusions (air bubbles). I suspect the glass is thick enough, but....
My concern about envelope temperature only relates to the few I have with inclusions (air bubbles). I suspect the glass is thick enough, but....
Yes, when you start going much about 500V, many things change- ask yourself why sweep tubes either use plate caps or have isolated plate pins. Arcing and corona become serious issues as any ham who has built high power linears will tell you.
The wrist strap will not help you- that's useful for static, not anything with current.
The wrist strap will not help you- that's useful for static, not anything with current.
I have had a few issues with corona, have not had arcing issues.
NOS GM70 I have purchased usually rustle and crackle around the base when first installed. Quite disconcerting the first time you encounter this. A few hours of use usually solves this, sometimes (rarely) a good scrub with alcohol and a toothbrush helps.
All wiring in my amps carrying a kV or more is double insulated. (Note at full power the output transformer primary sees 1kVdc and as much 1kVpp in class A1 operation, and was designed for these operating voltages - well for A2 actually.)
Note that I have NEVER worked on these amps with 1kV applied to the output stage. Covers on and all measurements done externally. An accident is so permanent and I am not willing to give myself the benefit of the doubt.
When I was younger I worked on ham gear with some pretty high voltages (not this high) and did not worry about it.
Other designs utilized raw supplies of up to 700V, even there I saw quite a difference as compared to what I see at 1kV.
NOS GM70 I have purchased usually rustle and crackle around the base when first installed. Quite disconcerting the first time you encounter this. A few hours of use usually solves this, sometimes (rarely) a good scrub with alcohol and a toothbrush helps.
All wiring in my amps carrying a kV or more is double insulated. (Note at full power the output transformer primary sees 1kVdc and as much 1kVpp in class A1 operation, and was designed for these operating voltages - well for A2 actually.)
Note that I have NEVER worked on these amps with 1kV applied to the output stage. Covers on and all measurements done externally. An accident is so permanent and I am not willing to give myself the benefit of the doubt.
When I was younger I worked on ham gear with some pretty high voltages (not this high) and did not worry about it.
Other designs utilized raw supplies of up to 700V, even there I saw quite a difference as compared to what I see at 1kV.
As a ham radio operator for decades, I've worked on a lot of RF amplifiers. Once worked on one with 7500v on the plate.
I got shocked once on a 1800v amp. I was shorting the caps to ground before servicing it and the voltage jumped through the insulation on the screwdriver, ran through my arm and popped open a 1/4" opening on my elbow...
Be careful
Mark
I got shocked once on a 1800v amp. I was shorting the caps to ground before servicing it and the voltage jumped through the insulation on the screwdriver, ran through my arm and popped open a 1/4" opening on my elbow...
Be careful
Mark
A wrist strap for working with HT supplies should be maid of braided wire and have a good connection to ground – (I have one that is about a ½" wide and looks like the ground strap of a car battery) – the rubber ones have more resistance than your skin and aren't even "useful for static" (IMHO).
The point I was making earlier is that one shouldn't be afraid of 1000V and unafraid of 500V - both can kill you. Always work with one hand in your pocket; always discharge caps†; never work "live" unless you absolutely have to (and even then, make sure you clip your probes on with power off first, then switch power on, take your reading, switch power off, and wait for the voltage to drop back to a safe level before you remove your test clips).
†Include bleeder resistors on all caps in the HT supply. If working on a device that doesn't have them, build yourself a "discharge rig." My high-voltage one is a 100K 25W resistor attached across a spare set of 2500V-rated DVM probes with an alligator clip on the black lead. I clip that to ground, carefully touch the other to the cap to be discharged, and wait until the meter drops below 20V. Then I wait a few minutes and repeat to allow for dielectric absorption – (it's not that the DA "rebound" will reach lethal volts, but you don't want a "surprise spark" causing you to flinch and wreck something).
Yes, but the 1000V one can kill you in more ways, some of which are surprising (for the millisecond before you die).
Never a bad idea in any amp, but for those that don't know about it:
Pomona 6733 hookup wire:
18ga w/10KV silicone insulation – < $25 a (50') roll and worth every penny for this sort of work
My GM70 SE runs on 1100v and I've never experienced the pops or cracks that Kevin describes.
I'm using Pomona wire and everything is carefully insulated, I've even epoxy potted as many exposed parts as I could. Others parts are varnished. I've also rounded any sharp metal part that carries HV.
I'm using Pomona wire and everything is carefully insulated, I've even epoxy potted as many exposed parts as I could. Others parts are varnished. I've also rounded any sharp metal part that carries HV.
I had the biggest problem with the copper version of the GM70, the graphite plate ones on hand seem to settle down quickly and permanently.
I have some upgrades planned for the driver stage, and when I do this I will take a file to the corners of the plate pin solder tab as you did. Everything else has large quantities of air around it.
No plans at this time to make the design publicly available.
I have some upgrades planned for the driver stage, and when I do this I will take a file to the corners of the plate pin solder tab as you did. Everything else has large quantities of air around it.
No plans at this time to make the design publicly available.
NB, Wrist strap.
If you use a wrist strap connected to earth while working on HV equipment. (don't do it!)
You are guaranteeing a good contact shock and will probably not survive.
You must try and isolate yourself from all contact with earth. You are Trying to create no connection with your body and any form of earth.
Chassis earth is one example. The other is any return path to the equipment supply. That means if the amp is disconnected from earth the chassis will be connected (internally)to the 0V side of supply and is just as dangerous.
Think like this, how do birds sit on power lines? The answer is the bird is at the same potential as the power line..ie there is no flow through the bird except the resistance between its feet. What would happen if the bird was connected to ground? BANG NO BIRD!
Regards
M. Gregg
If you use a wrist strap connected to earth while working on HV equipment. (don't do it!)
You are guaranteeing a good contact shock and will probably not survive.
You must try and isolate yourself from all contact with earth. You are Trying to create no connection with your body and any form of earth.
Chassis earth is one example. The other is any return path to the equipment supply. That means if the amp is disconnected from earth the chassis will be connected (internally)to the 0V side of supply and is just as dangerous.
Think like this, how do birds sit on power lines? The answer is the bird is at the same potential as the power line..ie there is no flow through the bird except the resistance between its feet. What would happen if the bird was connected to ground? BANG NO BIRD!
Regards
M. Gregg
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.