Hello - I've been around for a bit but don't post much. This DIY hobby is largely responsible for helping me maintain most of my sanity in my off hours during these trying times, but recently it's having the opposite effect.
My current project is my most ambitious to date, and essentially follows the schematic found Thorsten's article "Calculating DC coupled Single Ended Valve Amplifiers", specifically post #12 (https://www.diyaudio.com/forums/tub...d-single-valve-amplifiers-1-a-post372236.html).
The power supply is housed in a separate chassis, feeding DC for the heaters, filaments and B+. Chassis ground is separate from circuit ground and is common to both chassis. The 5842s share regulated 6.3VDC, and each 45 has its own DC supply and Coleman regulator.
Overall, the amp appeared to work properly. While I had sanity checked the supply voltages and dialed in the filaments, and did some initial listening through test speakers, I didn't do any further measurements. However, after some listening in the main rig it was clear that the right channel was louder than the left. Not drastically, but enough that I noticed pretty quickly. Swapping driver and power tubes and interconnects had no effect; the right channel was always the louder of the two. I measured the cathode current, and the right channel was lower by 10-15mA. I then discovered that the voltage drop across the OPTs was not consistent - there was a roughly 20% discrepancy. They are Lundahls, which require some jumpering by the consumer, and I have since confirmed that they were both wired correctly.
Then things got weird.
The resistance across both the primary and secondary was consistent between both OPTs. I also confirmed that there was no continuity between primary and secondary, nor between circuit ground and chassis ground. The negative binding posts are both connected to circuit ground, and confirmed with a meter. However, I found that there was continuity (1 ohm or so; a dead short is 0.7 ohms on my cheapo multimeters) between the L positive binding post and the R negative binding post, as well as between the R and L positive binding posts!
I have never seen this before, and I'm at my wits end having traced wiring and checking for continuity at various places. Though there are a few new twists for me with this particular build, in general this is not a complicated circuit - but I am stumped. I don't see how both OPTs could be coupled this way, let alone still appear to operate more or less correctly in this state.
I'm sure it's something simple that I'm overlooking, and it wouldn't be the first time, but if anyone can point me in the right direction I would be forever grateful.
Cheers,
Joe
My current project is my most ambitious to date, and essentially follows the schematic found Thorsten's article "Calculating DC coupled Single Ended Valve Amplifiers", specifically post #12 (https://www.diyaudio.com/forums/tub...d-single-valve-amplifiers-1-a-post372236.html).
The power supply is housed in a separate chassis, feeding DC for the heaters, filaments and B+. Chassis ground is separate from circuit ground and is common to both chassis. The 5842s share regulated 6.3VDC, and each 45 has its own DC supply and Coleman regulator.
Overall, the amp appeared to work properly. While I had sanity checked the supply voltages and dialed in the filaments, and did some initial listening through test speakers, I didn't do any further measurements. However, after some listening in the main rig it was clear that the right channel was louder than the left. Not drastically, but enough that I noticed pretty quickly. Swapping driver and power tubes and interconnects had no effect; the right channel was always the louder of the two. I measured the cathode current, and the right channel was lower by 10-15mA. I then discovered that the voltage drop across the OPTs was not consistent - there was a roughly 20% discrepancy. They are Lundahls, which require some jumpering by the consumer, and I have since confirmed that they were both wired correctly.
Then things got weird.
The resistance across both the primary and secondary was consistent between both OPTs. I also confirmed that there was no continuity between primary and secondary, nor between circuit ground and chassis ground. The negative binding posts are both connected to circuit ground, and confirmed with a meter. However, I found that there was continuity (1 ohm or so; a dead short is 0.7 ohms on my cheapo multimeters) between the L positive binding post and the R negative binding post, as well as between the R and L positive binding posts!
I have never seen this before, and I'm at my wits end having traced wiring and checking for continuity at various places. Though there are a few new twists for me with this particular build, in general this is not a complicated circuit - but I am stumped. I don't see how both OPTs could be coupled this way, let alone still appear to operate more or less correctly in this state.
I'm sure it's something simple that I'm overlooking, and it wouldn't be the first time, but if anyone can point me in the right direction I would be forever grateful.
Cheers,
Joe
Each secondary winding has a low DC resistance between its 16R and its common,
and each common is grounded.
So either positive output would be low ohms to either negative (grounded) output, through the ground.
(Or also between the two 16R outputs.)
Monoblocks would not do this (with the AC cords unplugged), unless the inputs were connected,
which would connect the grounds of the two chassis together, back at the source.
and each common is grounded.
So either positive output would be low ohms to either negative (grounded) output, through the ground.
(Or also between the two 16R outputs.)
Monoblocks would not do this (with the AC cords unplugged), unless the inputs were connected,
which would connect the grounds of the two chassis together, back at the source.
Last edited:
Of course - it's just series resistance! How silly of me.
What threw me is that I checked on another amp, and there is no continuity between the positive posts. However, I may not have the negatives tied to ground - so apples and oranges.
What threw me is that I checked on another amp, and there is no continuity between the positive posts. However, I may not have the negatives tied to ground - so apples and oranges.
Some tube amps do not connect the secondaries to ground, but this is a very, very bad practice, and
dangerous. They can only do this if feedback is not used, when an output ground reference is not necessary.
dangerous. They can only do this if feedback is not used, when an output ground reference is not necessary.
Last edited:
Why is it dangerous? Is it because of a possible short between the primary and secondary, with the result that high voltage could appear on the loudspeaker(terminals)?
There is an exception. In the times of AC-DC amplifiers (no mains transformer) the rule was: Do not 'ground' the secondary because of the hazard of electric shock through the loudspeaker(terminals). For feedback a separate winding on the OPT was necessary. This extra winding could than safely be connected to 'ground'.
Chasing ground connections/loops is a pain, as is figuring out earth loops. I spent a happy hour trying to figure out why my amps ground was connected to mains earth after I'd unplugged everything I could see and un-soldered half the amp. After a while you end up muttering to yourself, scratching your head and doubting your own sanity, bit like dropping acid or a belladona/gimpson weed cocktail but without the pretty colours.
Andy.
Andy.