I keep the old populated boards in a big box for reasons like this. It's far easier to fix up an old board with a few new parts when a situation like this arises. In the case of the mosfets in the TSE and TSE-II it has been a never ending parts chase from day one, as mosfets come and go quickly with some having a production lifetime of less than a year. before the 10M45 existed I had been using some "constant current diodes" from Motorola when the Motorola sales engineer told me that they were just selected mosfets with their gate and source tied together. This led to several mosfet experiments and before I came to a useful solution, the 10M45 appeared and the first gen TSE was born. There have been periods of time when 10M45's got scarce before, but the DN2540 was available and works fine with a resistor change.
I found my bag of DN2540N5's, stuck one in the board, and it does work as expected needing a slightly smaller resistor to make the same current as a 10M45. Some early morning experiments on the complete signal chain suggests that best overall THD can be achieved with 6 to 8 mA, but further testing is needed. I need to bust out a TSE or TSE-II board for testing as well. TSE board number 1 is still alive and well in my Lexan TSE amp. The TSE-II board will not fit in that tiny amp so it will stay all original until it dies.
Unfortunately the DN2540N5 is virtually non-existent these days. I like the looks of the IXTP1R6N100D2. Spec sheet says 1,000v's X 1 amp =100 watts. It will be running somewhere in the 300v or less range at around 10ma in the Tubelab boards. Price is comparable to the 10M45s and it should survive with the voltage surge on startup being rated for1,000v. Only true test is put some in and put some hours on them.
The DN2535N5 is rated for 350 volts. The DN2540N5 is rated for 400 volts. These are otherwise identical parts, so the DN2535N5 is likely a grade out from the wafer that did not meet the 400 volt spec with some manufacturers margin. This means that the 2535 part can exhibit breakdown conditions between 350 and 400 volts. There is no part rated for more than 400 volts so the DN2540 could breakdown at 401 volts or it could withstand far more voltage. There is no way to tell other than testing.
The B+ voltage at start up in an SSE can venture well over 450 volts depending on the transformer, rectifier and line voltage. At initial startup the 12AT7 is not conducting and therefore drawing no current from the CCS part. There will be zero voltage across the part at start up. As the 12AT7 warms up it will begin to draw current which puts voltage across the CCS part and the 10K resistor. Once the 12AT7 reaches a temperature where it tries to draw more than the CCS set current (7 to 10 mA) the CCS will increase its voltage drop to keep the current at the set current value. At a set current of 7 mA there will be 70 volts dropped across the 10 K resistor and 150 to 300 volts across the 12AT7 leaving less than 300 volts across the CCS part. The real surge comes from a loud music transient, or a brief interruption in line power. I have used a pair of DN2540N5s in a TSE for well over a year with no issues but have not tried them in the SSE until yesterday.
The DN2535N5 should work, but I would spend the extra 2 bucks and get some IXTP1R6N100D2s. They have been hammered pretty hard in my SSE for 3 days with no issues, but 3 days testing in one amp is not a valid guarantee of success either. It is however a 1000 volt part. I'm going to try the IXTP08N50D2 next. It's a little cheaper.
SSE board #1 has been on my bench for a few days. Before ripping it off the bench and putting it back into "the box" I decided to revisit the experiment that literally made the biggest bang, but on a smaller scale. I have known for some time that one could drive the two channels out of phase, then wire a P-P OPT across the two plates to turn the SSE into a not so Simple PP. I made a few guitar amps this way many years ago. Trying this with TV sweep tubes scattered electrolytic goo all over my workroom when I turned the power supply up a wee bit too much.
I used a small Edcor interstage matching transformer for a phase inverter and wired a 6600 ohm guitar amp OPT across the plates with the tubes jumpered for pure pentode operation. With a B+ voltage of over 500 volts I saw 50 watts at about 3% THD flow from the SSE board using the Blues Tubes 6L6WGT's. I tried some "expendable" Chinese KT88's from the 1990's, but lost ground due to excess grid current. I rolled through a bunch of driver and output tubes and a few different OPT's with varying degrees of success, but my goals are not my typical "MORE POWER." I need to find a driver design capable of feeding a BIG cathode follower output stage. Power is not important, I need CLEAN voltage.
I stumbled across a combination that got my attention. It works too good to be true but it is. Good enough that I want to dig up another dead SSE board and build a second copy for listening in stereo, but I only have one OPT. What's so special? How about a little amp that only makes 10 watts from a pair of 6L6's in class A push pull. These numbers were seen on two separate occasions yesterday, and verified this morning. The residual THD in the audio oscillator is 0.04%. The 100 mW reading is affected by residual hum and noise from the sloppy wiring.
POWER THD
0.1W 0.103%
0.5W 0.056%
1W 0.068%
2W 0.1265%
5W 0.328%
10W 0.488%
12W 1.984% onset of clipping
15W 5.40% serious symmetrical clipping with slight rounding
The OPT is an old style Hammond 1608 wired for 8000 ohms. Tubes are the Chinese "Blues Tubes" 6L6WGBs and the driver is a JAN 6189W which is a 12AU7 that doesn't suck.
I used a small Edcor interstage matching transformer for a phase inverter and wired a 6600 ohm guitar amp OPT across the plates with the tubes jumpered for pure pentode operation. With a B+ voltage of over 500 volts I saw 50 watts at about 3% THD flow from the SSE board using the Blues Tubes 6L6WGT's. I tried some "expendable" Chinese KT88's from the 1990's, but lost ground due to excess grid current. I rolled through a bunch of driver and output tubes and a few different OPT's with varying degrees of success, but my goals are not my typical "MORE POWER." I need to find a driver design capable of feeding a BIG cathode follower output stage. Power is not important, I need CLEAN voltage.
I stumbled across a combination that got my attention. It works too good to be true but it is. Good enough that I want to dig up another dead SSE board and build a second copy for listening in stereo, but I only have one OPT. What's so special? How about a little amp that only makes 10 watts from a pair of 6L6's in class A push pull. These numbers were seen on two separate occasions yesterday, and verified this morning. The residual THD in the audio oscillator is 0.04%. The 100 mW reading is affected by residual hum and noise from the sloppy wiring.
POWER THD
0.1W 0.103%
0.5W 0.056%
1W 0.068%
2W 0.1265%
5W 0.328%
10W 0.488%
12W 1.984% onset of clipping
15W 5.40% serious symmetrical clipping with slight rounding
The OPT is an old style Hammond 1608 wired for 8000 ohms. Tubes are the Chinese "Blues Tubes" 6L6WGBs and the driver is a JAN 6189W which is a 12AU7 that doesn't suck.
Attachments
Just for reference, I have some 10M90S, out of 10m45s, and being curious I decided to use my Peak Atlas DCA75 to see what it says about the part. It shows up as a N ch depletion mosfet on this device. What made me check is buying some 10M45S from China and curious to see what the Peak said about them. The Chinese 10M45S' showed as N ch enhancement mosfet. Not even a good fake. Buyer beware of Chinese offerings of these parts. Money spent was refunded when told they were fake though.
There was some talk here:
https://www.diyaudio.com/community/threads/tubelab-se-removing-mosfets.291593/post-4887819
on making the SSE, into a fast push pull amp, for anyone that wants to go this route.
I've been meaning to try this - have the dual R core power transformers, some OPT's with tertiary taps for sweep tubes, and a big solid copper case for the whole shebang, and a plan for making it into a dual monoblock amp, with FM stereo tuner.
Just never seem to make the time to move it forward in the project backlog.
Win W5JAG
https://www.diyaudio.com/community/threads/tubelab-se-removing-mosfets.291593/post-4887819
on making the SSE, into a fast push pull amp, for anyone that wants to go this route.
I've been meaning to try this - have the dual R core power transformers, some OPT's with tertiary taps for sweep tubes, and a big solid copper case for the whole shebang, and a plan for making it into a dual monoblock amp, with FM stereo tuner.
Just never seem to make the time to move it forward in the project backlog.
Win W5JAG
Here is another little tidbit of information that came in handy way back in my teens and pre-teenage years. Back then my parts depot for building stuff was the local landfill or trash dump. The place was filled with old TV B&W sets, radios and occasionally a console HiFi set. Every time I visited, I took some tools and retrieved as many tubes as I could carry and maybe a transformer or two.
A ham radio guy who was the older brother of a classmate and I took apart a 5C1 Fender Champ and traced its schematic so I could build one, or a dozen. 5Y3's and 6SJ7's were pretty common in old radios but 6V6's were rather scarce. 6K6's were more common and worked pretty good, but I discovered that the 6BQ6GT and 6DQ6 sweep tubes will plug right into the 6V6 socket if you wire a plate cap robbed from a TV set to pin 3. Do NOT do this on a stock Champ. You will fry the power transformer down the road!
A 6BQ6GA or 6BQ6GT will work in an SSE if you keep the screen voltage below 200 volts, or total B+ voltage down into the 300 volt range. They really shine in the push pull version with well over 100 watts flowing before meltdown. A 6DQ6 can be made to work in the SEE, but careful tweaking and some higher voltage rated cathode bypass caps are needed. This is how the electrolytic goo got spread all over the SSE boards, and why it still has 100 volt bypass caps in it.
I've built 2 SPP's and will build another soon, I think its a great amp but what I'd really want to build next is an EL34 or 6L6-GC PP, so definitely interested the octal PP if it happens.
Thank very much for taking the time to write about your experimentation and trials!
Is this a trustworthy source for 10M45S?
10M45S Ebay
Is this a trustworthy source for 10M45S?
10M45S Ebay
I would not trust any coming from China from my experience. Just get IXTP1R6N100D2. Mosfet's have been used to make a CCS for years.
I putchased some off ebay. 2 out of the 10 were bad, the rest functioned but had a wide range of variation. I ended up testing each one under close to operating conditions to find a matched pair and adjusting resistors to suit the application. My guess is these did not pass quality during manufacture and ended up on ebay, buyer beware.
The 332ohm resistor will determine the current running through the tube. You want less than 10ma.
