Was poking around a bit and found the tab on the Q102 and Q202 mosfets has continuity with ground. I am using the non-isolated FQP9P25. First I thought it might be an issue with the mounting but even when I pull them away from the heatsink I get continuity. I'll give the board another once over.
Been doing some more measuring and assuming Ohms law still hold up 
The current I(R106) + I(R107) = I(R105), in my case a little over 1mA or 8V/7k5 on g1 of the driver pentode.
However, I(R107) = I(R108), so no current goes through the tube, no matter the position of the pot.
The 180V and 52V supplies are ok.
12HG7 or EF184 makes no difference.
The pot goes from 0-24V, also on the wiper into R103.
The other side of R103 however will go up ok, but won't go much below 10V.
With g1 at 8V, the pentode is still at cutoff.
With the wiper at 0 and 10V at the other end of R103, that voltage must come from somewhere: the VP0106 must be leaking.
The current I(R106) + I(R107) = I(R105), in my case a little over 1mA or 8V/7k5 on g1 of the driver pentode.
However, I(R107) = I(R108), so no current goes through the tube, no matter the position of the pot.
The 180V and 52V supplies are ok.
12HG7 or EF184 makes no difference.
The pot goes from 0-24V, also on the wiper into R103.
The other side of R103 however will go up ok, but won't go much below 10V.
With g1 at 8V, the pentode is still at cutoff.
With the wiper at 0 and 10V at the other end of R103, that voltage must come from somewhere: the VP0106 must be leaking.
I unplugged the driver tubes and did the first two tests, after a little warm up I got 195V on the screen grid and 56V on D4. I tried connecting the meter to pin 9 but could not get a reading.
Spiggs,
Measuring the B+ across R1 is correct.
310 volts on the plate of the driver tube and 0 volts on the cathode says that the tube is not conducting any current. What tube type are you using? Are you feeding it the correct heater voltage and does it get warm? I don't see any obvious problems with your board build.
Parafeed813,
Make sure that your mosfet is installed correctly. The picture that Spiggs posted is correct.
I have never tried to run my board from 280 volts. My testing has been in the 375 to 600 volt range. I'll try a lower voltage as soon as I can, maybe tomorrow.
Measuring the B+ across R1 is correct.
310 volts on the plate of the driver tube and 0 volts on the cathode says that the tube is not conducting any current. What tube type are you using? Are you feeding it the correct heater voltage and does it get warm? I don't see any obvious problems with your board build.
Parafeed813,
Make sure that your mosfet is installed correctly. The picture that Spiggs posted is correct.
I have never tried to run my board from 280 volts. My testing has been in the 375 to 600 volt range. I'll try a lower voltage as soon as I can, maybe tomorrow.
That was it, not the incorrect voltage but no voltage to the heaters. Fixed the wiring and I was able to dial in the plate voltage. Stupid me I was focused on the meters and not the fact that the heaters were cold. I am using 6J51P tubes for the drivers so it's the same 6.3V source as I am using for the outputs.
Onto the output tubes now.
Success! I bias the output tube up to about 20mA and let them sit an hour while keeping an eye on the meters. Bias did move around for about 15min then stabilized. After an hour I turned up the bias to 45mA with the B+ now showing 490V. Let the amp sit awhile then shut it down. Hooked up a music source and speakers started it back up and heard music. No critical listening of course but my old Realistic speakers sounded as good as they ever have.
Attachments
Last edited:
Music
Replaced the small three legged creatures and the EF184 driver pentodes came to life.
I set the plate voltage to around 130V (which takes some fiddling with a regular pot, next time a multiturn).
Pop in some EL36, let them wake up slowly at 20mA and after a while put them to work at 50mA.
These are '12W' sweeps, so 50mA @ 275V (-40V cathode) is safe.
Far from Tubelab territory still, but I did turn the knob up, readjusting to 50mA. 275-350-400V.
Power output into 4ohms was around 5W no matter the voltage.
Into 8ohms (6k5) it went from under 4W to 7½W at 350V. 400V was slightly lower. No serious (or accurate) testing yet.
Music first...
Replaced the small three legged creatures and the EF184 driver pentodes came to life.
I set the plate voltage to around 130V (which takes some fiddling with a regular pot, next time a multiturn).
Pop in some EL36, let them wake up slowly at 20mA and after a while put them to work at 50mA.
These are '12W' sweeps, so 50mA @ 275V (-40V cathode) is safe.
Far from Tubelab territory still, but I did turn the knob up, readjusting to 50mA. 275-350-400V.
Power output into 4ohms was around 5W no matter the voltage.
Into 8ohms (6k5) it went from under 4W to 7½W at 350V. 400V was slightly lower. No serious (or accurate) testing yet.
Music first...
Good progress; I’m just ordering parts
Good to know some Beta Board builders are listening to music already - well done!
I’ve been away on vacation for several weeks and have just now put together a parts order.
Ordering from Mouser (USA) presented several hurdles, and several parts are in low supply, if not out of stock. I’m noting some here in case it may help others.
1. C2 as listed on George’s BOM is out of stock, now actually marked “Not Stocked” - I ordered the 150uF item in the same series - MAL215919151E3 of which 400 are currently in stock. https://www.mouser.com/ProductDetai...KYqh34kqWbg==&countrycode=US¤cycode=USD
2. C101, C201 listed as FKP3F022204C00JSSD is a Non-stock item. I ordered instead MKP1G022203C00KS. https://www.mouser.com/ProductDetai...zJaTAPg/05Q==&countrycode=US¤cycode=USD
3. I’m not sure I found the right pots. George indicated that the adjustment potentiometers are generating only 100 milliwatts, but I could not find any 100k multi-turn pots that will fit the PCB. I ordered these: http://www.mouser.com/ProductDetail...35wNKcgKTUQ==&countrycode=US¤cycode=USD
I hope the future “real Tubelabs PCB for UNSET” will provide holes for .5 watt multi-turn pots. I have a few more suggestions regarding the PCB and fitment of parts, but I will save them for later after I some fun and soldered my PCB.
Good to know some Beta Board builders are listening to music already - well done!
I’ve been away on vacation for several weeks and have just now put together a parts order.
Ordering from Mouser (USA) presented several hurdles, and several parts are in low supply, if not out of stock. I’m noting some here in case it may help others.
1. C2 as listed on George’s BOM is out of stock, now actually marked “Not Stocked” - I ordered the 150uF item in the same series - MAL215919151E3 of which 400 are currently in stock. https://www.mouser.com/ProductDetai...KYqh34kqWbg==&countrycode=US¤cycode=USD
2. C101, C201 listed as FKP3F022204C00JSSD is a Non-stock item. I ordered instead MKP1G022203C00KS. https://www.mouser.com/ProductDetai...zJaTAPg/05Q==&countrycode=US¤cycode=USD
3. I’m not sure I found the right pots. George indicated that the adjustment potentiometers are generating only 100 milliwatts, but I could not find any 100k multi-turn pots that will fit the PCB. I ordered these: http://www.mouser.com/ProductDetail...35wNKcgKTUQ==&countrycode=US¤cycode=USD
I hope the future “real Tubelabs PCB for UNSET” will provide holes for .5 watt multi-turn pots. I have a few more suggestions regarding the PCB and fitment of parts, but I will save them for later after I some fun and soldered my PCB.
I ended up ordering from Mouser, Digikey, and Newark to find all the bits excluding the tube sockets and transformers.
So I let the UNSET run a good 10 hours straight today and all it did was make music so all seems good. Next up I am putting together a soft start board and will wire it up a bit better so I can ditch the clip leads I am currently using for grounding I also would like to setup a way to test the THD so I can do some fine tuning. Long term plan is to put the UNSET in a nice case and replace the SSE I currently use in my main system.
Question George. What kind of watts would you expect from my setup with a B+ of 490V using 6DQ5 output tubes into 3K Toroidy OPTs? Am I in the ballpark targeting 45mA of bias?
Honestly I am surprised by the B+ readings using the Antek 3T350. On my SSE using a 2T350 I get a B+ of 430V. I am using a solid state rectifier for both. Is it just the extra capacity of the 3T version?
spiggs, good news! A few quick questions for your. What driver tube are you using, what trimmer pots did you get?
I will have a similar transformer setup, except that I’ll be using an Antek 4T320 as power transformer. So I expect B+ of ~450V. What is the voltage across your 6DQ5 (anode to cathode)? I guess about 440V, so running 45 ma gives a plate dissipation under 20 watts.
Any observations on the musical qualities of your UNSET? Compared to SSE?
I will have a similar transformer setup, except that I’ll be using an Antek 4T320 as power transformer. So I expect B+ of ~450V. What is the voltage across your 6DQ5 (anode to cathode)? I guess about 440V, so running 45 ma gives a plate dissipation under 20 watts.
Any observations on the musical qualities of your UNSET? Compared to SSE?
Last edited:
I am using the Amphenol Piher PT10MV10-104A2020-S for my trimmer pots. Driver tubes are 6J51P.
I jumped on the UNSET beta build because I thought it would be a good way to push myself to learn more about tube amps. The only other I have built is the SSE. So I am a bit ignorant on where all the proper places to measure are. For the voltage across the tube where is the right place to pick this up? I did measure at the end of r115 and r215 and see 458V there. This is also why I am being conservative on the bias until I understand better what I should be watching.
As far as sound I am very happy with what I hear so far. These speakers are some old Optimus (Realistic/Radio Shack) Pro X7s. I used these to test the SSE as well and usually listen to them in my office with a Neurochrome Modulus-86 amp. The UNSET drives them well. It has the SE sound that made me want another tube amp after the SSE but seems to have more authority and some good punch lower down. It will play the simple stuff but did not shy away when I asked it to rock. Not that the SSE did but this has a little more.
I jumped on the UNSET beta build because I thought it would be a good way to push myself to learn more about tube amps. The only other I have built is the SSE. So I am a bit ignorant on where all the proper places to measure are. For the voltage across the tube where is the right place to pick this up? I did measure at the end of r115 and r215 and see 458V there. This is also why I am being conservative on the bias until I understand better what I should be watching.
As far as sound I am very happy with what I hear so far. These speakers are some old Optimus (Realistic/Radio Shack) Pro X7s. I used these to test the SSE as well and usually listen to them in my office with a Neurochrome Modulus-86 amp. The UNSET drives them well. It has the SE sound that made me want another tube amp after the SSE but seems to have more authority and some good punch lower down. It will play the simple stuff but did not shy away when I asked it to rock. Not that the SSE did but this has a little more.
I posed a similar question before, so I try to answer myself to see if in the meanwhile I've found the right answer.
45 mA on 3k means 135 V more to reach zero volts, and 135 V less to reach twoce the current. If I've understood correctly, I'd expect to use a 3k load on 45 mA at around 185 V B+ (that is 50 + 135) and swing up to 320V. But this with a tube with 7 W of plate dissipation.
6DQ5's plate dissipation is 24 W, so 490 V and 45 mA give around 20 W on the tube only, and that is fine.
Waiting for George's confirmation, at 490 V 45 mA I would use a 10k load, so that it swings from 40V up to 940 V with around 10-11 Wrms.
With 5k load around 400V and 73 mA that give around 12-13 Wrms. 320 V at 92 mA could give around 12-13 Wrms on 3k load. With 2k load I'd go towards 270 V and 110 mA with around 11-12 Wrms, etc...
Am I right?
Thanks
Let the UNSET run for another 10 hours straight today. No smoke or funny smells, just music. I have meters hooked up to B+, output tube bias, and driver tube plate voltage. A few observations.
My initial B+ reading of 490V is the highest I have seen, it fluctuates with the lowest reading at 475V so far. Typically it has been hovering around 477-480V. Is this just mains voltage fluctuation? Would it be of value to add a regulator to the B+ and set it at a lower but stable voltage?
I agree with Parafeed813, those one turn pots take some fiddling to dial in. If i am targeting say 45mA I can get to 45ish (between 44.9 to 45.9) but very small movements make big changes.
The bias drifts around with the changes in B+, how far out is too much, both overall and imbalance between channels? Is 1-3 mA higher lower or imbalance too much? This is what I have been seeing when checking on it throughout the day while playing music.
The amp is quiet when no music is playing, I have not put my ear right up next to the speakers yet but from a foot away I hear nothing when no music is playing. There is some microphonics when adjusting the pots. Considering this is all grounded with clip leads and with cheap meters hooked up using more clip leads I think this speaks well for ultimately how quiet the amp will be.
Finally and most important I like listening to it!
My initial B+ reading of 490V is the highest I have seen, it fluctuates with the lowest reading at 475V so far. Typically it has been hovering around 477-480V. Is this just mains voltage fluctuation? Would it be of value to add a regulator to the B+ and set it at a lower but stable voltage?
I agree with Parafeed813, those one turn pots take some fiddling to dial in. If i am targeting say 45mA I can get to 45ish (between 44.9 to 45.9) but very small movements make big changes.
The bias drifts around with the changes in B+, how far out is too much, both overall and imbalance between channels? Is 1-3 mA higher lower or imbalance too much? This is what I have been seeing when checking on it throughout the day while playing music.
The amp is quiet when no music is playing, I have not put my ear right up next to the speakers yet but from a foot away I hear nothing when no music is playing. There is some microphonics when adjusting the pots. Considering this is all grounded with clip leads and with cheap meters hooked up using more clip leads I think this speaks well for ultimately how quiet the amp will be.
Finally and most important I like listening to it!
Let the UNSET run for another 10 hours straight today. No smoke or funny smells, just music. I have meters hooked up to B+, output tube bias, and driver tube plate voltage. A few observations.
My initial B+ reading of 490V is the highest I have seen, it fluctuates with the lowest reading at 475V so far. Typically it has been hovering around 477-480V. Is this just mains voltage fluctuation? Would it be of value to add a regulator to the B+ and set it at a lower but stable voltage?
I agree with Parafeed813, those one turn pots take some fiddling to dial in. If i am targeting say 45mA I can get to 45ish (between 44.9 to 45.9) but very small movements make big changes.
The bias drifts around with the changes in B+, how far out is too much, both overall and imbalance between channels? Is 1-3 mA higher lower or imbalance too much? This is what I have been seeing when checking on it throughout the day while playing music.
The amp is quiet when no music is playing, I have not put my ear right up next to the speakers yet but from a foot away I hear nothing when no music is playing. There is some microphonics when adjusting the pots. Considering this is all grounded with clip leads and with cheap meters hooked up using more clip leads I think this speaks well for ultimately how quiet the amp will be.
Finally and most important I like listening to it!
My initial B+ reading of 490V is the highest I have seen, it fluctuates with the lowest reading at 475V so far. Typically it has been hovering around 477-480V. Is this just mains voltage fluctuation? Would it be of value to add a regulator to the B+ and set it at a lower but stable voltage?
I agree with Parafeed813, those one turn pots take some fiddling to dial in. If i am targeting say 45mA I can get to 45ish (between 44.9 to 45.9) but very small movements make big changes.
The bias drifts around with the changes in B+, how far out is too much, both overall and imbalance between channels? Is 1-3 mA higher lower or imbalance too much? This is what I have been seeing when checking on it throughout the day while playing music.
The amp is quiet when no music is playing, I have not put my ear right up next to the speakers yet but from a foot away I hear nothing when no music is playing. There is some microphonics when adjusting the pots. Considering this is all grounded with clip leads and with cheap meters hooked up using more clip leads I think this speaks well for ultimately how quiet the amp will be.
Finally and most important I like listening to it!
This is one of the main reasons that UNSET exists. TV sweep tubes are less costly that 300B's or most KT88's, but they have a far higher peak current handling capability for slapping non cooperative speakers around when they need it.
First off 500 volts from an Antek 3T350 is a bit high. I get about 475 volts under load from a 4T360 with 122 volts of line voltage. Check the batteries in your red volt meters. I know that the older yellow Harbor Freight meters will read high when the batteries get weak. I haven't run the red pair I recently got down low enough to tell if they do the same.
490 volts calls for something higher than a 3K load impedance, since 3K really wants a lot more than 45 mA. I was getting around 20 watts into 3K on 100 mA, but I was using 26HU5's and only 400 volts.
Those who simulate and calculate could tell us what to expect from any combination of tubes and transformers, but for me it's just easier to fire up my amp.
I have a 1500 ohm Toroidy OPT in there now, so I simply put my 8 ohm load on the 4 ohm tap to reflect a 3K load to the tubes. I have a 26HU5 tube in place, and I'm not motivated to mess with the wires, so it stays. Most sweep tubes are within a few percent of each other when operated well below their peak current ratings, which is the case here. The big tube also allows me to explore power levels beyond what a small tube can do. Any reasonable horizontal sweep tubes cam come within a few percent of these numbers within it's dissipation specs.
To take these measurements I used an old HP8903A that can read THD and "watts into X ohms."
To determine tube dissipation 3 voltage readings and the current reading are needed. Black lead on ground and red lead on B+ gives you the B+ voltage. Move the red lead to the end of r115 and r215. This is the plate voltage. Then move the red lead to the bus line across the board marked cathode for the tube that you are making measurements on.
To determine the tube dissipation you subtract the cathode voltage from the plate voltage, then multiply the result by the bias current. SO:
With 490 volts of B+, I set the bias to 45 mA and this gives 485 volts on the plate with my OPT. I measure 74 volts on the cathode, so the tube has 485 - 74 = 411 volts across it. 411 volts times .045 amps is 18.495 watts. I measure 5.5 watts of output under these conditions at the edge of clipping.
Running a 3K ohm load on nearly 500 volts puts your tube into its dissipation limit too early. With 490 volts of B+ I hit 24 watts at 58 mA. that should get you about 9 to 10 watts of power. Your OPT will have a slightly higher DCR, so you will have a bit less plate voltage.
Running slightly over spec at 70 ma and 28.8 watts of dissipation gives 14 watts of power at clipping.
70 mA with 400 volts of B+ gives 13 watts of power, but the tubes are cooler at 23.7 watts of dissipation. 350 volts seems to be the sweet spot for a 24 watt tube, producing 12.2 watts of power with 21 watts of dissipation in the tube.
Variations in line voltage will cause changes in B+ voltage and tube current, as with any unregulated tube amp, or an old TV set where these tubes were used. A few percent is not a problem, ten percent or more is a problem. I live out in the country where everybody has electric heat. Some of my amps need to be turned up in the winter when my line voltage drops 5 to 8 volts.
I suspect that your B+ will come down a bit with more current, but your tubes will be happier and you will get a bit more power with a lower B+ or a higher load impedance.
I made a chart of what a 3K OPT can do under several different B+ voltages and tube dissipations. I will test other impedances and post the results in the next few days.
Your 3K ohm OPT will look like a 6K ohm OPT if you run an 8 ohm speaker on the 4 ohm tap. Some OPT's work well this way, some don't. The only way to find out is to try it. Making no other changes try listening to music with the speakers on each tap. see which sounds better to you with the music you like most.
I have some 3K Transcendar OPT's in my SSE. I run them at 6K most of the time by plugging my speakers into the 4 ohm jacks. Most music has better bass and transient response this way at a slight loss in power. With the extra B+ you might even gain power this way.
You could run a bit more current if your B+ was reduced. There are a few ways to do this. A tube rectifier is the easiest, but not so easy with an Antek transformer.
Dear George,
I've done a simple excel file:
- first column is the max plate dissipation (choose the one of your specific tube);
- first raw defines the percentage of max plate dissipation the whole system (tube + pmosfet) runs at (95% should be reasonable in this configuration, ad the tube alone should run around 85% of max plate dissipation).;
- second raw is the voltage at which the amp will work;
- the number in the specific cell is the optimal load to swing the plate simmetrically down to 50 V (g2 to be set accordingly).
I've done a simple excel file:
- first column is the max plate dissipation (choose the one of your specific tube);
- first raw defines the percentage of max plate dissipation the whole system (tube + pmosfet) runs at (95% should be reasonable in this configuration, ad the tube alone should run around 85% of max plate dissipation).;
- second raw is the voltage at which the amp will work;
- the number in the specific cell is the optimal load to swing the plate simmetrically down to 50 V (g2 to be set accordingly).
Attachments
Oh, the adventures of Beta board building without a proper BOM. My Mouser order arrived and I have to fess up; screwed up with points 2 and 3 above:
2. C101, C201 should be WIMA FKP3G022204D00KJ00 to fit the 15 mm pitch. Mouser has them:
FKP3G022204D00KJ00 WIMA | Mouser
3. The Amphenol/pt15rv15-104a2020 part I ordered does not fit the PCB! I’m keeping my fingers crossed and reordering this one now:
https://www.mouser.com/ProductDetai...worthy/409HS-100K?qs=GWr7qGobaGO8xYW1ocYUSg==