Hello everyone,
I've been trying to fix this issue on my GM70 amp for half a year without success, i'm now turning to the diyaudio community as i'm completely out of ideas on what it could be (hopefully something simple as usual 😀)
So it goes like this: The Amp works fine with one GM70 inserted (doesn't matter which one in which socket) and the bias can be cranked up to 100mA and above. It does fluctuate very slightly (+/- 0.5v) which can be seen on the analogue meters i got on the front (3V meter across 10R resistor in cathode) but it never starts going out of control and i can listen to music fine. Now as soon as the second GM70 is inserted and bias on that cranked up the bias on both tubes starts oscillating wildly going from 0-2V on the front panel meters. This happens roughly when a total cathode current of 100mA is reached. Cranking both tubes to 30-40mA somewhat works, although some slight oscillation can already be seen on the meters. When cranked further it goes completely out of control, at 45mA it starts self-oscillating so that it becomes worse and worse without me touching the bias pots.
Here is a schematic of my Bias supply, i copied it from here apart from the Rectifier and Filter.
Edit: Just noticed C4/C5 are connected in reverse in this schematic, they are the other way round in reality.
And here is the schematic of the Amp itself:
Further observations and what i already tried:
More powerful Bias Transformer: No change
Removing 10µF Caps at the bias supply outputs: No change
Measured B+ with Oscillation present: Barely fluctuates (900V +/- 5V, with 50mA Bias per tube which somewhat keeps oscillation under control between 10 and 120mA so not to stress the tubes too much)
Measured Bias Voltage with Oscillation present: Barely fluctuates (+/- 1V)
Measured DC Offset on Grids: None apart from the Bias Voltage
Measured Driver Stage by itself, without GM70s inserted: Gives me clean 180Vpp on both outputs without oscillation
So this makes me think it's some sort of interaction between the tubes, but where are they connected? Could the two Bias "Legs" interfere with each other somehow? Output Transformers are connected correctly, at least according to Toroidy's Datasheet. Secondary Ground is connected to Amp Ground.
Pics of the Amp can be found here, not much has changed since then apart from OT's being installed and some wires rerouted.
I've been trying to fix this issue on my GM70 amp for half a year without success, i'm now turning to the diyaudio community as i'm completely out of ideas on what it could be (hopefully something simple as usual 😀)
So it goes like this: The Amp works fine with one GM70 inserted (doesn't matter which one in which socket) and the bias can be cranked up to 100mA and above. It does fluctuate very slightly (+/- 0.5v) which can be seen on the analogue meters i got on the front (3V meter across 10R resistor in cathode) but it never starts going out of control and i can listen to music fine. Now as soon as the second GM70 is inserted and bias on that cranked up the bias on both tubes starts oscillating wildly going from 0-2V on the front panel meters. This happens roughly when a total cathode current of 100mA is reached. Cranking both tubes to 30-40mA somewhat works, although some slight oscillation can already be seen on the meters. When cranked further it goes completely out of control, at 45mA it starts self-oscillating so that it becomes worse and worse without me touching the bias pots.
Here is a schematic of my Bias supply, i copied it from here apart from the Rectifier and Filter.
Edit: Just noticed C4/C5 are connected in reverse in this schematic, they are the other way round in reality.
And here is the schematic of the Amp itself:
Further observations and what i already tried:
More powerful Bias Transformer: No change
Removing 10µF Caps at the bias supply outputs: No change
Measured B+ with Oscillation present: Barely fluctuates (900V +/- 5V, with 50mA Bias per tube which somewhat keeps oscillation under control between 10 and 120mA so not to stress the tubes too much)
Measured Bias Voltage with Oscillation present: Barely fluctuates (+/- 1V)
Measured DC Offset on Grids: None apart from the Bias Voltage
Measured Driver Stage by itself, without GM70s inserted: Gives me clean 180Vpp on both outputs without oscillation
So this makes me think it's some sort of interaction between the tubes, but where are they connected? Could the two Bias "Legs" interfere with each other somehow? Output Transformers are connected correctly, at least according to Toroidy's Datasheet. Secondary Ground is connected to Amp Ground.
Pics of the Amp can be found here, not much has changed since then apart from OT's being installed and some wires rerouted.
The bias current is minimal , one tube , two , it doesn't matter .
The power supply schematic is important , there is a big change when you use both tubes . I don't see much filtering caps in picture
The power supply schematic is important , there is a big change when you use both tubes . I don't see much filtering caps in picture
Likely the main power supply capacitors have degraded in capacitance value, causing low frequency instability.
Or else it was always there.
Post a schematic of the main power supply. This is a monoblock amplifier?
Or else it was always there.
Post a schematic of the main power supply. This is a monoblock amplifier?
How are you powering the cathodes? Is there interaction there? Is the double power drawn with 2 tubes (instead of 1) causing some issues?
Depanatoru and rayma,
B+ does not sag significantly, however there may be some margin of error here if it fluctuates faster than my multimeter can measure.
PSU uses 650V 0.3A transformer into 1.6KV bridge -> CC->L->C->L->C. Caps are 40µF 1.1KV Epcos DC Link. Chokes are Hammond 159V installed in the ground leg. PSU ground is connected to amp ground after the 2nd choke ofc.
ErikdeBest,
Each GM70 has it's own 19.5V 10A SMPS (from Dell workstation Notebook). Grounds are floating.
B+ does not sag significantly, however there may be some margin of error here if it fluctuates faster than my multimeter can measure.
PSU uses 650V 0.3A transformer into 1.6KV bridge -> CC->L->C->L->C. Caps are 40µF 1.1KV Epcos DC Link. Chokes are Hammond 159V installed in the ground leg. PSU ground is connected to amp ground after the 2nd choke ofc.
ErikdeBest,
Each GM70 has it's own 19.5V 10A SMPS (from Dell workstation Notebook). Grounds are floating.
You should try to short the two chokes and with just all capacitors in parallel to see how it work
the 300mA may be marginal for 200mA DC, but should be more than sufficient to feed each valve 45mA without problems, so I do not think that is the problem.
When you say the grounds of the SMPS are floating, you mean the gnd pin on the 230VAC input to the PS? Did you try other PS? Did you try powering both tubes from the same SMPS (you can use the voltage drop over the RDC of the OPT to check for current). I think the 0V of the DC side of the SMPS may be connected to the ground?...
When you say the grounds of the SMPS are floating, you mean the gnd pin on the 230VAC input to the PS? Did you try other PS? Did you try powering both tubes from the same SMPS (you can use the voltage drop over the RDC of the OPT to check for current). I think the 0V of the DC side of the SMPS may be connected to the ground?...
No the Primary Earth Pin/PE is connected to Amp Ground which is also connected to Mains Earth. Secondary Ground is connected to Amp Ground via the 10R resistor.
Did i misunderstand DC Heating DHT's and PE needs to be floating? No idea how this is connected internally, they are enclosed units.
I did not try powering both tubes from the same SMPS, i can give it a shot tomorrow. Unfortunately i have no other supplies at hand that can provide 20V (or close to it) at 3A.
Did i misunderstand DC Heating DHT's and PE needs to be floating? No idea how this is connected internally, they are enclosed units.
I did not try powering both tubes from the same SMPS, i can give it a shot tomorrow. Unfortunately i have no other supplies at hand that can provide 20V (or close to it) at 3A.
This is likely a low frequency phase margin problem, not just DC voltage sag.
Add more capacitance in parallel with C3 in diagram in post #5 (at least 100uF more) and see if the problem is reduced.
You may need to add even more than that.
Add more capacitance in parallel with C3 in diagram in post #5 (at least 100uF more) and see if the problem is reduced.
You may need to add even more than that.
How are the GM70 paralleled ? Or are you talking the 2 tubes of a stereo setup? If paralleled they may be generating RF oscillations against one another.
rayma, i will see what i can find in terms of additional caps and try that. This guy seems to use the same configuration but with a single 10H choke without issue tho.
Edit: Nvm, it's only one tube per channel.
Potentiallyincorrect, i'm talking about the 2 tubes in a stereo amp yes. One per channel.
Edit: Nvm, it's only one tube per channel.
Potentiallyincorrect, i'm talking about the 2 tubes in a stereo amp yes. One per channel.
There can be a connection between PE and the ground/negative of the DC connector. If that is the case, you would be shorting out the 10R resistor. Should theoretically not lead to oscillations, but there will be some ground loops and these are never good. I would therefore start by connecting the cathodes of both valves in parallel and see what happens - someone against this?
Short the chokes to see quickly if the problem is there , after that we can think what should be done
No free lunch.
Poorly designed PSU == swings.
Redesigning is required.
p.s.
Are you sure, that all capacitors -at least- 1200V, high ripple current type?
The first capacitor's charging current is few A when empty!!!
Poorly designed PSU == swings.
Redesigning is required.
p.s.
Are you sure, that all capacitors -at least- 1200V, high ripple current type?
The first capacitor's charging current is few A when empty!!!
Last edited:
Erik, in fact there most likely is a connection PE-GND. These were used for Notebooks with metal cases.. so definitely something to try.
Depanatoru, will give that a shot as well tomorrow.
euro21, so more capacitance? I cannot observe this extreme sag when measuring and the voltage is higher than in the simulation. 900V with a single tube running at 100mA, sim shows 840.
Depanatoru, will give that a shot as well tomorrow.
euro21, so more capacitance? I cannot observe this extreme sag when measuring and the voltage is higher than in the simulation. 900V with a single tube running at 100mA, sim shows 840.
At least short out L21, if you don't have more capacitors to parallel with the last capacitor.
That alone may help enough.
That alone may help enough.
Maybe he could move one 40uF from input to output ...
disso , the voltage is not sagging , the inductance and capacitance have such values that is oscillating . Of course adding more capacitance is always the solution . But if you don't have more capacitors maybe could be rearranged .
disso , the voltage is not sagging , the inductance and capacitance have such values that is oscillating . Of course adding more capacitance is always the solution . But if you don't have more capacitors maybe could be rearranged .
Space is also an issue, but may be able to fit one more cap in there. Or go smaller stacked electrolytics.
Caps are Epcos B32778G0406, rated at 21.5A RMS ripple current. They are 1100V, not 1200.
Caps are Epcos B32778G0406, rated at 21.5A RMS ripple current. They are 1100V, not 1200.
If you increase C2 and C3 (for example up to 100-120uF), the swinging will decrease.
If you want to use electrolytics, this is the working 211 PSU (voltage doubler).
If you want to use electrolytics, this is the working 211 PSU (voltage doubler).