Well hard to say what stopper best stops a lumped parasitic oscillation?
Those that interact locally, only within the circuit itself... A resistor is
not going to "stop" a phase margin problem or design error of that sort...
But if its interacting with free space (radio), thats much easier to say!
Mono and dipoles will have an impedance between 25 to 75 ohms cross
the feedpoint, and maybe 450 to 800 ohms (to GND) at far open ends.
Impedance of free space is 377ohms. I don't know if that last figure is
actually relevant in this argument?
So the point I make: is that a 1 ohm stopper is not going to significantly
spoil the efficiency of a parasitic antenna stub, nor is 10K doing anything
but breaking the problem into two separate resonant stubs... Capacitively
coupled right across the stopper from one High Z stub end to another...
If the stop resistance is not between 25 < R < 600, I don't think it stands
much chance for spoiling the Q of the stub or killing RF.
If higher resistance has to be used for some other reason, then stick a
smaller value in series on either side to spoil the two new child stubs.
A string of small value resistors is probably more effective than any one
large lump resistor of equal sum.
Last edited:
Thanks Kenpeter. I think my problem was the dodgy leads that I made up for my test setup. Everything appears OK now. I will try monitoring the output when hooked up to the main system again just to be sure that I don't have a systemic problem with the amp layout.
Tubelab, thanks again for your support and assistance. I will hook up balance adjustment as you describe. The trim pots in my parts bin are 10 turn types, so no hair trigger on adjustment. I think I will also try reducing the plate load resistors in the diff amps to up the current in them to about 7mA each. Thanks for the tip regarding the software, will investigate!
Also, when initially looking at design setup, I thought 400 volt B+ would be suitable as I was looking at 6L6GC, 6BG6GA and 7591. My test tubes are 6P3S. They don't like 400 volts (I had a Tubelab moment!)
Cheers,
Chris
Tubelab, thanks again for your support and assistance. I will hook up balance adjustment as you describe. The trim pots in my parts bin are 10 turn types, so no hair trigger on adjustment. I think I will also try reducing the plate load resistors in the diff amps to up the current in them to about 7mA each. Thanks for the tip regarding the software, will investigate!
Also, when initially looking at design setup, I thought 400 volt B+ would be suitable as I was looking at 6L6GC, 6BG6GA and 7591. My test tubes are 6P3S. They don't like 400 volts (I had a Tubelab moment!)
Cheers,
Chris
For the balance control, does this look like what is described?
Thanks,
Chris
Uploaded with ImageShack.us
Thanks,
Chris
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
Whilst waiting for parts I have been doing some thinking/tinkering. My heater voltage lift is only set for +55 volts. I will re-set for +75 volts for better heater to cathode limit compliance. On power up now, watching the voltages, the -ive supply does not pull it out of limits, so don't think I need any clamping device as suggested in past.
Looking at the 7-8mA 6SN7 likes for linearity, I was quoting Morgan Jones. With some re-reading, I have seen that he has about 5mA per triode in his diff amp stage of his Crystal Palace amp, as the diff amp cancels second order harmonic distortion. I decided to do what I could to see any distortion without any analysis equipment. I hooked up my cheap and simple battery operated signal generator, removed the output tubes and placed the scope probes on the input and output. Inverse on channel 2 of the scope and placed one waveform over the other to see any differences.
First scope picture is at 1kH square wave, output superimposed over input. Not too bad!
Next sine wave at 1kHz
Here is sine wave at 20kHz. Some phase shift.
Did not have square wave at 20kHz due to limitations of cheap signal generator. Overall, pretty happy with results and don't feel I need to tweek too much, maybe just change the 47K first stage load resistors for 33k and get the current up just a wee bit.
Still waiting on parts to implement the balance circuit above. Maybe this weekend.
Looking at the 7-8mA 6SN7 likes for linearity, I was quoting Morgan Jones. With some re-reading, I have seen that he has about 5mA per triode in his diff amp stage of his Crystal Palace amp, as the diff amp cancels second order harmonic distortion. I decided to do what I could to see any distortion without any analysis equipment. I hooked up my cheap and simple battery operated signal generator, removed the output tubes and placed the scope probes on the input and output. Inverse on channel 2 of the scope and placed one waveform over the other to see any differences.
First scope picture is at 1kH square wave, output superimposed over input. Not too bad!
An externally hosted image should be here but it was not working when we last tested it.
Next sine wave at 1kHz
An externally hosted image should be here but it was not working when we last tested it.
Here is sine wave at 20kHz. Some phase shift.
An externally hosted image should be here but it was not working when we last tested it.
Did not have square wave at 20kHz due to limitations of cheap signal generator. Overall, pretty happy with results and don't feel I need to tweek too much, maybe just change the 47K first stage load resistors for 33k and get the current up just a wee bit.
Still waiting on parts to implement the balance circuit above. Maybe this weekend.
Last edited:
Power supply failure
Had the opportunity to work on the amp for a few hours today. After some research reading MJ, I saw that he had similarly low current flow through his diff amps in the Crystal Palace amp, noting that the diff amp naturally cancels the second order harmonic distortion. Therefore, I will keep values as suggested by Tubelab.
I was in the process of replacing one resistor in the heater lift divider (old schematic had about 55Volts, new value is about 80volts) when I blew my power supply. The resistor is on the power supply board. After replacing it and re-mounting the board, I powered up and noticed I was getting no voltage. Some initial trouble shooting revealed that the 100R 10W resistor in the input C R C was open circuit. A little odd, but I had mounted it high on its leads to give it some cooling room, and the constant jiggling around as I was building may have damaged something. I replaced it and powered up. Got a good 396 volt reading (what I was getting previously), for about 3 seconds... Then BANG! A bright flash and I killed the power. there was some evidence of arcing from the tab of the power MOSFET to the chassis. Figuring it was probably blown, I replaced it. Powered up, and getting about 260 volts output. Hmm, measured adjust to out on the LM317 and it was about 0.8 volts, so thought this may have gone out in sympathy, so replaced it too. Re-powered and same result.
With all tubes removed, I get about 430 volts out of the supply. With the 2 6SN7s in place, I get about 260 volts. Raw DC at the input filter caps is around 540 volts. I measure 260 at the source and drain of the 2SK1119 MOSFET and at the output of the LM317, and about 0.8v between OUT and ADJ on the LM317.
Any suggestions about how to trouble shoot this? What is likely to have gone out in sympathy when the 2SK1119 arced to ground?
Unfortunately I am about to step out the door and will be away for a couple of days with work, so no response to any suggestions does not indicate a lack of appreciation.
Regards,
Chris
Schematic of power supply:
http://img171.imageshack.us/img171/855/powersupply.pdf
Had the opportunity to work on the amp for a few hours today. After some research reading MJ, I saw that he had similarly low current flow through his diff amps in the Crystal Palace amp, noting that the diff amp naturally cancels the second order harmonic distortion. Therefore, I will keep values as suggested by Tubelab.
I was in the process of replacing one resistor in the heater lift divider (old schematic had about 55Volts, new value is about 80volts) when I blew my power supply. The resistor is on the power supply board. After replacing it and re-mounting the board, I powered up and noticed I was getting no voltage. Some initial trouble shooting revealed that the 100R 10W resistor in the input C R C was open circuit. A little odd, but I had mounted it high on its leads to give it some cooling room, and the constant jiggling around as I was building may have damaged something. I replaced it and powered up. Got a good 396 volt reading (what I was getting previously), for about 3 seconds... Then BANG! A bright flash and I killed the power. there was some evidence of arcing from the tab of the power MOSFET to the chassis. Figuring it was probably blown, I replaced it. Powered up, and getting about 260 volts output. Hmm, measured adjust to out on the LM317 and it was about 0.8 volts, so thought this may have gone out in sympathy, so replaced it too. Re-powered and same result.
With all tubes removed, I get about 430 volts out of the supply. With the 2 6SN7s in place, I get about 260 volts. Raw DC at the input filter caps is around 540 volts. I measure 260 at the source and drain of the 2SK1119 MOSFET and at the output of the LM317, and about 0.8v between OUT and ADJ on the LM317.
Any suggestions about how to trouble shoot this? What is likely to have gone out in sympathy when the 2SK1119 arced to ground?
Unfortunately I am about to step out the door and will be away for a couple of days with work, so no response to any suggestions does not indicate a lack of appreciation.
Regards,
Chris
Schematic of power supply:
http://img171.imageshack.us/img171/855/powersupply.pdf
Last edited:
Well Russ, I had replaced the pass FET and the 317 previously and tested with bad result.
I tested all of the resistors, and they checked out OK. One of the protection zeners tested bad, so I figured that it probably failed it's protection duty after replacing the FET and 317. Taking your advice regarding expected failures, I bit the bullet and replaced all actives (FET, 317 and diodes). Maybe overkill, but the few cents in parts offset the time involved if I were to keep blowing things.
Result - power supply all fixed!
I implemented the balance circuit described by Tubelab. The result was not what I expected. Could not get a stable voltage. I took the back of the envelope sketch down to the garage from amongst the pile of crap on my desk, turns out I had taken the wrong circuit diagram down.
My way at the moment, one step forward, two steps back!
It is up and running at least, without the balance control. I will work on that over the weekend and then she should be complete and ready to build its twin. Here are some shots of the scope.
Triangle wave trace 1 input trace 2 output of MOSFET drive, just backed off from clipping at about 130 volts from 1.2 volts input.
Uploaded with ImageShack.us
This is the square wave result at 10kHz. Think Ch1 is output of MOSFET, Ch2 is output transformer with 6R dummy load.
Just getting to know my new toy (Rigol 50mHz digital scope, software hacked to 100mHz).
Uploaded with ImageShack.us
I tested all of the resistors, and they checked out OK. One of the protection zeners tested bad, so I figured that it probably failed it's protection duty after replacing the FET and 317. Taking your advice regarding expected failures, I bit the bullet and replaced all actives (FET, 317 and diodes). Maybe overkill, but the few cents in parts offset the time involved if I were to keep blowing things.
Result - power supply all fixed!
I implemented the balance circuit described by Tubelab. The result was not what I expected. Could not get a stable voltage. I took the back of the envelope sketch down to the garage from amongst the pile of crap on my desk, turns out I had taken the wrong circuit diagram down.
My way at the moment, one step forward, two steps back!
It is up and running at least, without the balance control. I will work on that over the weekend and then she should be complete and ready to build its twin. Here are some shots of the scope.
Triangle wave trace 1 input trace 2 output of MOSFET drive, just backed off from clipping at about 130 volts from 1.2 volts input.
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
This is the square wave result at 10kHz. Think Ch1 is output of MOSFET, Ch2 is output transformer with 6R dummy load.
Just getting to know my new toy (Rigol 50mHz digital scope, software hacked to 100mHz).
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
Last edited:
Regarding the new toy, I thought it might replace my basic old Tek 2225 50mHz scope that has no bells or whistles. Well, I won't be replacing it. I find the screen resolution of the Rigol quite poor (no surprise) compared to the Tek. For instance, I also purchased a new (cheap) function generator. It produces a sine wave that has point on the peak like a Norman arch. Very easy to see on the old Tek without any zooming in, not so easy on the Rigol. Where the new Rigol excels though, is having all the measurements right before your eyes and being able to store the screen shots to a USB stick. Many other useful features on the new scope that I am sure I will get to know and love, but it will still be sitting next to the old. BTW, when I was looking to get one, I thought that I should get a 100mHz scope rather than a 50, I found out that the 50 and 100 Rigol are actually the same machine, with a software switch to change model. Can be done with a USB stick and a 9 pin serial cable and a few terminal commands. Like you, I got a new SG as well. It does not have as nice a sine wave as my very cheap 'pocket audio oscillator', but the square wave is much better.
BTW, I can't believe I spent half a day putzing around putting in the balance circuit Tubelab suggested, finding it did not work, ripping it out, then finding out I had simply implemented it incorrectly
Anyway, it will give me something to play with over the weekend
Cheers
BTW, I can't believe I spent half a day putzing around putting in the balance circuit Tubelab suggested, finding it did not work, ripping it out, then finding out I had simply implemented it incorrectly
Anyway, it will give me something to play with over the weekend
Cheers
When going down a path that not as often traveled 1 step forward, 2 steps back is to be expected. You really need to try and avoid the wrong turns and the 1 step forward 5 steps back. My experimenting time has been very scarce lately since I am stuck at work a lot. In the past few months I have blown the red board up again, removed the mosfet experiments and restored it to the original function (2 steps forward, 2 back, but knowledge gained). I made some progress with 3 different Simple P-P builds (all forward), and dragged my 6L6GC AB2 breadboard back onto the bench. Here some forward progress was made before I decided to turn down a side street. It is beginning to look like this was not a wrong turn. Time for experiments is scarce since I often don't get home from work until 9PM, but this is what has happened.
The breadboard was dragged back onto the bench, and wiring was completed for operation in triode mode with 6L6GC's. All running on bench power supplies. A few pictures are still in my camera. No measurements were made but I had it running with a CD player and speakers for about 2 weeks. It sounded very nice.
If you notice my tone throughout the first 5 years of my presence on these pages you will notice my preference for triode operation with NO feedback. Petes red board, and the Simple P-P has changed my opinion quite a bit. I had 4 different P-P amps up and running, and 3 of them are pentode mode with some type of feedback. I can't decide which one I like the most, but I can say that the 6CW5 based Simple P-P is my least favorite just because it makes 12 WPC while the others are all over 50 WPC. There is something to be said about the "effortlessness" of a big amp at 1/4 throttle, compared to a little guy at full tilt. I have not listened to a SE amp in a few months.
Notice I said "had". There are still 4 working amps, but as of now none of them are triode mode. The 6L6GC amp is now running with the 6L6GC's in pentode mode with Schade style feedback. The input tube is now a 6SL7 to provide a bit more gain. It flat rocks the house and I really need to stop and take some measurements and pictures. More tweaking is needed, but I will stop and document what I did when I have time.
We all do that, including me. That is still better than spending half a day wiring stuff up only to have the fire gods dance all over it before you can shut it off. I got this monster HP power supply that is really cool for testing big stuff since it goes up to 650 volts at 1.7 amps. Unfortunately it shows no mercy when you screw up. It will scatter components and blow traces clean off your PCB's.
They replaced my 20 year old scope at work with a "budget" HP digital a few years back. It has lots of cool features, but the 8 bit resoultion does suck sometimes, and you really need to understand how aliasing happens before you can learn to trust what you see. I use a TEK 2232 at home, but it is beginning to fail (reboots often). I don't know what I will get when it dies though.
Ha Ha... Finally, after a year and a half I almost have one monoblock completed and ready to build the next, now there is something else worth trying with the design
I guess once the power supply and chassis are completed experimenting with these other things is pretty easy.
Next big decision is if I should be using these amps to drive my speakers full range or split tweeter duty to my Baby Huey (already have the active crossovers). If I do use the BH, I have to fix a ringing problem that I suspect is the output transformers. It does lead to more experimenting and iron purchases though
Thanks for the invaluable help George!
Chris
I guess once the power supply and chassis are completed experimenting with these other things is pretty easy.
Next big decision is if I should be using these amps to drive my speakers full range or split tweeter duty to my Baby Huey (already have the active crossovers). If I do use the BH, I have to fix a ringing problem that I suspect is the output transformers. It does lead to more experimenting and iron purchases though
Thanks for the invaluable help George!
Chris
Try it full range first, you may be suprised. My Opus (very similar to yours) was only for bass when I designed and built it. It will stay fullrange for quite some time as it does an incredible job (better than any amp I have built or purchased) across the board. I will eventually go bi- or tri amp just to do it, but the challenge will be how to build a better amp to to take over the duties for the mid and tweeters.
Ha Ha... Finally, after a year and a half I almost have one monoblock completed and ready to build the next, now there is something else worth trying with the design
I guess once the power supply and chassis are completed experimenting with these other things is pretty easy.
Next big decision is if I should be using these amps to drive my speakers full range or split tweeter duty to my Baby Huey (already have the active crossovers). If I do use the BH, I have to fix a ringing problem that I suspect is the output transformers. It does lead to more experimenting and iron purchases though
Thanks for the invaluable help George!
Chris
I wonder if a step circuit as described in "Amplifier Resonse & Stability and All That" (John Moyle) would be effective in addressing the ringing problem.
The article is on Pete Millett's web site.
Yes, I would love to try it full range first. My speaker crossovers have been disconnected and are tricky to get to... At the moment I have an old Adcom 555 driving the sub-100Hz through a stereo pair of woofer boxes, the n Baby Huey driving the mid range of my Ariels and a small 6P14P SE amp driving the tweeters.
Here is the active crossover using Rod Elliot's (Elliot Sound Products) boards with Linkwitz Transform circuit for bass drivers.
Uploaded with ImageShack.us
Here is the active crossover using Rod Elliot's (Elliot Sound Products) boards with Linkwitz Transform circuit for bass drivers.
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
Is this normal?
G'Day Folks,
Well, after what feels like 2 steps forward, three steps back, I have made some progress. I had some power supply issues. Seemed like every time I removed the power supply board to get to something, when I put it back in it would blow. I spent some time trying to work out what was going on and build a jig to test LM317 chips, as replacing every active component was getting expensive. I think the 317s were blowing from shorting to the chassis, even though it was attached with insulator and insulated insert for bolt. I have replaced the steel bolt with a plastic one and (fingers crossed) it appears working.
Next problem was high frequency 'whistle' that would sound when the external hard drive of my media PC was working. Initially thought it was an oscillation problem, but a hint from Wavebourn on another thread lead to resolving it was a ground loop problem. Solved with the standard back to back diodes//resistor//capacitor joining circuit ground to chassis.
Now, the amp appears to be working OK, making music, low distortion etc. The driver stage developed by Tubelab appears to be working well. My dumb question for the day is: No matter what bias I set, the output of the amp is the same. I am using a set of 6P3S-E tubes as my testers. I set the bias by measuring the voltage across the 1 ohm resistors to ground from the cathodes of the output tubes. If I set about 40mA idle current (can't remember bias voltage, but it is -30 something volts) I get a max output of 13.7 volts RMS across an 8 ohm dummy load, giving 23 Watts. The 13.7 volts RMS output coincides with the plates of the output tubes reaching clipping of 600 volts (observed on scope). Now, if I change the bias, nothing changes, max output at 13.7 VRMS across dummy resistor when plates clip at 600V. This is the same from about 10mA (where the mosfet driver clips, probably due to -42V bias eating in to the -105 supplied to the MOSFET) to 80mA, max output 13.7VRMS due to limit of 600V from power tubes plates. I tried different valves, with similar results. KT88 and 6L6 gave max 665V at plates for 14.6 VRMS = 27 Watts, EL34 allowed a 680 volt swing at plates for 15.3VRMS and 29 Watts. Have I screwed something up? I would have expected that the hotter I biased the tubes, the more output power I would have gained?
Attached is schematic if it helps,
Regards,
Chris
Uploaded with ImageShack.us
G'Day Folks,
Well, after what feels like 2 steps forward, three steps back, I have made some progress. I had some power supply issues. Seemed like every time I removed the power supply board to get to something, when I put it back in it would blow. I spent some time trying to work out what was going on and build a jig to test LM317 chips, as replacing every active component was getting expensive. I think the 317s were blowing from shorting to the chassis, even though it was attached with insulator and insulated insert for bolt. I have replaced the steel bolt with a plastic one and (fingers crossed) it appears working.
Next problem was high frequency 'whistle' that would sound when the external hard drive of my media PC was working. Initially thought it was an oscillation problem, but a hint from Wavebourn on another thread lead to resolving it was a ground loop problem. Solved with the standard back to back diodes//resistor//capacitor joining circuit ground to chassis.
Now, the amp appears to be working OK, making music, low distortion etc. The driver stage developed by Tubelab appears to be working well. My dumb question for the day is: No matter what bias I set, the output of the amp is the same. I am using a set of 6P3S-E tubes as my testers. I set the bias by measuring the voltage across the 1 ohm resistors to ground from the cathodes of the output tubes. If I set about 40mA idle current (can't remember bias voltage, but it is -30 something volts) I get a max output of 13.7 volts RMS across an 8 ohm dummy load, giving 23 Watts. The 13.7 volts RMS output coincides with the plates of the output tubes reaching clipping of 600 volts (observed on scope). Now, if I change the bias, nothing changes, max output at 13.7 VRMS across dummy resistor when plates clip at 600V. This is the same from about 10mA (where the mosfet driver clips, probably due to -42V bias eating in to the -105 supplied to the MOSFET) to 80mA, max output 13.7VRMS due to limit of 600V from power tubes plates. I tried different valves, with similar results. KT88 and 6L6 gave max 665V at plates for 14.6 VRMS = 27 Watts, EL34 allowed a 680 volt swing at plates for 15.3VRMS and 29 Watts. Have I screwed something up? I would have expected that the hotter I biased the tubes, the more output power I would have gained?
Attached is schematic if it helps,
Regards,
Chris
An externally hosted image should be here but it was not working when we last tested it.
Uploaded with ImageShack.us
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.