Also could someone explain to me what physically happens when two two tubes, sharing a cathode connection, crosstalk? I mean... when tube conducts it will draw current from the low impedance connection to gnd and not the other tube's cathode. Why would i see a signal on the other tube' plate? Sorry for the dum questions but i am learning as I go..
An externally hosted image should be here but it was not working when we last tested it.
Just did a complete high voltage stress test. Everything checks out. Ready to test some audio!
Just a quick question.... dht tube are fed through a gnd connection. Remove that connection and the tubes will stop conducting. What if i add a small resistor therby introducing some feedback within the output tube? The smallish voltage produced will affect bias (relative voltage differential seen by the grids)... does the idea have any merit at all?
If the tubes had separate filament supplies, the resistor would introduce negative feedback. Since they share a supply, they also must share a ground connection and a resistor, so it will also introduce crosstalk. Signal in one channel will bleed through to the other via the shared resistor.
There will also be RFI suppression capacitors inside the SMPS, between output terminals and chassis. These will effectively be in parallel with this resistor, bypassing it at high frequencies.
There will also be RFI suppression capacitors inside the SMPS, between output terminals and chassis. These will effectively be in parallel with this resistor, bypassing it at high frequencies.
Last edited:
There will be no significant crosstalk. Ground the negative end of the power supply, and the positive end should also look like a short to ground at audio frequencies, there is no path for signal to get from one channel to the other.
Of course, there is no "significant crosstalk" between channels of most amplifiers, but that doesn't stop people building dual monoblocks. When you're an audiophile, anything can be "significant" whether it exists or not.
Of course, there is no "significant crosstalk" between channels of most amplifiers, but that doesn't stop people building dual monoblocks. When you're an audiophile, anything can be "significant" whether it exists or not.
Thank you. Precisely the kind of fb i needed!
Btw had some issues with the tube rectifiers but my xenons are up and running!
They are in the gu46 screen ps section. Just for thr sake of the colours, no real practical needs. The only intetesting feature is that the tibes work with by359f diodes in a hybrid circuit. Save some transformer winding.
Btw had some issues with the tube rectifiers but my xenons are up and running!
They are in the gu46 screen ps section. Just for thr sake of the colours, no real practical needs. The only intetesting feature is that the tibes work with by359f diodes in a hybrid circuit. Save some transformer winding.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Burning the midnight oil...
Well,
some things work some things don't.... (usual)
the output section is very very stable. No trace of oscillation and a small signal fed onto the grid yields a very nice output.
But this is just the testing phase.
then I tested the driver and here is where things got worse.
I have a terrible motorboating issue which I suspect come from missing decoupling caps. The two halves of both stages are fed through a common voltage pole. I suspected it could happen with the first stage (200-230v 100k to plate of the 6sn7) but the use of the ccs on the second stage (6sl7 or 6bl7) suggested a different outcome.
well, here I am working on the driver... At this stage could anyone point to a working 2 stage all triode driver which can yield at least a 100vpp signal?
Thanks!
some things work some things don't.... (usual)
the output section is very very stable. No trace of oscillation and a small signal fed onto the grid yields a very nice output.
But this is just the testing phase.
then I tested the driver and here is where things got worse.
I have a terrible motorboating issue which I suspect come from missing decoupling caps. The two halves of both stages are fed through a common voltage pole. I suspected it could happen with the first stage (200-230v 100k to plate of the 6sn7) but the use of the ccs on the second stage (6sl7 or 6bl7) suggested a different outcome.
well, here I am working on the driver... At this stage could anyone point to a working 2 stage all triode driver which can yield at least a 100vpp signal?
Thanks!
Here's my suggestion:
1) move the CCS-loaded 1/2 6SL7 to the V1 position, and use 1/2 of a resistor-loaded 6BL7 as V2. This should swing plenty of volts.
2) As I posted earlier, transmitting tubes like to see low grid-ground impedance. I'd also recommend using a grid choke if possible - if not, try to keep the size of this grid resistor below 40k.
1) move the CCS-loaded 1/2 6SL7 to the V1 position, and use 1/2 of a resistor-loaded 6BL7 as V2. This should swing plenty of volts.
2) As I posted earlier, transmitting tubes like to see low grid-ground impedance. I'd also recommend using a grid choke if possible - if not, try to keep the size of this grid resistor below 40k.
Hello to all and Happy New Year.
The gu46 amp is proving quite a challenge. Here are the latest issues and improvements:
1. The PSU stuttering.
Cold filaments will cause a temporary overload. I noticed the phenomenon with the led switch on button. It will stutter. Eventually the PSU manages to heat the filaments but I fear the continuous strain will cause premature damage.
My solution was to place a 20A 1ohm thermistor in the secondary. Obviously a thermistor will not work in the primary of a switching supply so I must intervene on the secondary. I haven't had the chance to value impact on sonic performance and the added heat in the chassis is a liability.
So I have thought of bypassing the thermistor with a relay but implementation is proving challenging.
2. The driver section was replaced as a result of hum, motorboating, etc.
I have opted for a single stage per channel 12hg7 ala Pete Millet. Fully bypassed the pentode provides excellent gain and I can achieve full voltage swing over the entire anode voltage of the gu46 (70v RMS). Anode of the 12hg7 is circa 400v and the screen 150v. I can omit the bypass cap and reduce gain accordingly.
3. Decoupling.
this has proven a true challenge from an engineering standpoint. The chassis is small and already "drilled and painted". However I noticed that full decoupling (of every single line) is paramount for good performance. I therefore fully decoupled the driver stage (150v + 400vv) and the lines feeding into gu46 (1500v + 600v+ and -150v).
For the driver I opted for 600v 12uf polipropilene radial caps. 1500v and -150v lines have been decoupled using 10uf electrolytes (in series) whilst the 600v line with 5uf MKP caps.
Also the coupling caps on the driver have been upgraded to orange 630v 0.47uf non polarized caps.
4. Softstart
Anode lines need some time to fully turn on (cathode stripping + gas rectifier protection) so I implemented a SS circuit board. Very sturdy but even then a thermistor is in order to avoid blows to all rectifiers.
Well that is it for now. Hopefully this will help others. I have built large amps and thought I could get away with some of the more sophisticated solutions but I was wrong. Even 1W out of a transmitting tube requires enormous efforts to keep everything under control.
The gu46 amp is proving quite a challenge. Here are the latest issues and improvements:
1. The PSU stuttering.
Cold filaments will cause a temporary overload. I noticed the phenomenon with the led switch on button. It will stutter. Eventually the PSU manages to heat the filaments but I fear the continuous strain will cause premature damage.
My solution was to place a 20A 1ohm thermistor in the secondary. Obviously a thermistor will not work in the primary of a switching supply so I must intervene on the secondary. I haven't had the chance to value impact on sonic performance and the added heat in the chassis is a liability.
So I have thought of bypassing the thermistor with a relay but implementation is proving challenging.
2. The driver section was replaced as a result of hum, motorboating, etc.
I have opted for a single stage per channel 12hg7 ala Pete Millet. Fully bypassed the pentode provides excellent gain and I can achieve full voltage swing over the entire anode voltage of the gu46 (70v RMS). Anode of the 12hg7 is circa 400v and the screen 150v. I can omit the bypass cap and reduce gain accordingly.
3. Decoupling.
this has proven a true challenge from an engineering standpoint. The chassis is small and already "drilled and painted". However I noticed that full decoupling (of every single line) is paramount for good performance. I therefore fully decoupled the driver stage (150v + 400vv) and the lines feeding into gu46 (1500v + 600v+ and -150v).
For the driver I opted for 600v 12uf polipropilene radial caps. 1500v and -150v lines have been decoupled using 10uf electrolytes (in series) whilst the 600v line with 5uf MKP caps.
Also the coupling caps on the driver have been upgraded to orange 630v 0.47uf non polarized caps.
4. Softstart
Anode lines need some time to fully turn on (cathode stripping + gas rectifier protection) so I implemented a SS circuit board. Very sturdy but even then a thermistor is in order to avoid blows to all rectifiers.
Well that is it for now. Hopefully this will help others. I have built large amps and thought I could get away with some of the more sophisticated solutions but I was wrong. Even 1W out of a transmitting tube requires enormous efforts to keep everything under control.
Last edited:
