I have been digging into this more and while earlier in the thread I stated that it was stable - that is a small signal SPICE analysis. I agree with what you say but want to add that I believe that the amp goes unstable once cross-conduction starts and then self destructs even if the overload signal is removed that initially caused the overload. I've not tested this on the bench and it is just a theory for now. SPICE large signal time domain simulations of the amp in clipping did show large cross-conduction spikes as high as 3 amps as I recall.
I've finally come to the conclusion that the gain in the output stage causes the outputs to go deeper into saturation than with normal unity gain CFP or EF outputs and this is probably the reason for so many issues with this design.
Last edited:
I think we agreed some time ago that a Baker clamp on the driver stage would solve this issue - I have yet to try it in practice.
Most of the amps would blow up very nicely when merely plugging in or out of the RCA input jack, that was enough of a spike to fry the thing totally... some folks have reported that their units were stable, but that may be an artifact of the layout and wiring adding inductance or capacitance that slows the amp down, dunno...
_-_-
Most of the amps would blow up very nicely when merely plugging in or out of the RCA input jack, that was enough of a spike to fry the thing totally... some folks have reported that their units were stable, but that may be an artifact of the layout and wiring adding inductance or capacitance that slows the amp down, dunno...
_-_-
Pete,
I have not been able to "slow down" the amp to make it stable... thus far... I tried a few standard methods some years ago. I like the Baker clamp as shown a few posts ago, I think it is likely to solve the underlying problem.
Dunno, someone ought to try it.
_-_-bear
-- the problem appears to be that the driver (or the outputs) suffer from "shoot through" and staturate, lock in the "on" state, thus causing massive and fast overcurrent and device death... the Baker Clamp prevents this from happening, thus saving the amp. At least that is the idea... or so I am led to believe
I have not been able to "slow down" the amp to make it stable... thus far... I tried a few standard methods some years ago. I like the Baker clamp as shown a few posts ago, I think it is likely to solve the underlying problem.
Dunno, someone ought to try it.
_-_-bear
-- the problem appears to be that the driver (or the outputs) suffer from "shoot through" and staturate, lock in the "on" state, thus causing massive and fast overcurrent and device death... the Baker Clamp prevents this from happening, thus saving the amp. At least that is the idea... or so I am led to believe
Last edited:
I recently tried to revive some old Tiger .01s and a plastic Tiger I built long ago. Only one of the .01s work. I'll have to put the others on the bench and see what's up. They all prob need a recap.
I seem to recall a couple weird things about them...no on indicator lamp (just overtemp) and the peculiar external + and - rail fuses. I seem to recall if just one blew you got half the supply right across the speaker.
Can anyone suggest replacement output devices?
Anyway the one that works has been called back to duty in a microphone response test station (I design microphones).
Les
L M Watts Technology
I seem to recall a couple weird things about them...no on indicator lamp (just overtemp) and the peculiar external + and - rail fuses. I seem to recall if just one blew you got half the supply right across the speaker.
Can anyone suggest replacement output devices?
Anyway the one that works has been called back to duty in a microphone response test station (I design microphones).
Les
L M Watts Technology
IMHOWell, they are known to blow up real good, edgar. I was going to gut them and throw in a chip amp or something, but decided not to due to their historical perspective.
One thing I will throw in there though...a balanced input. A THAT line reciever should be fine. There's room for a TRS or XLR connector in the back I think. And it seems to already have a zener regulated +/- 17 volts
to power it.
Les
One thing I will throw in there though...a balanced input. A THAT line reciever should be fine. There's room for a TRS or XLR connector in the back I think. And it seems to already have a zener regulated +/- 17 volts
to power it.
Les
Mic's
Hi Les, I thought I recognized you from the Yahoo group!
Howdy Neighbor!!! Aaaah!, we'll need to have a good
conversation about those WM-61A's someday.
Hi Les, I thought I recognized you from the Yahoo group!
Howdy Neighbor!!! Aaaah!, we'll need to have a good
conversation about those WM-61A's someday.
Edgar,
you buy the beer and i'll talk microphones till i'm blue in the face!
The other Tiger .01 is fixed. It was just a bad connection. Hate to admit it, but I found a connection to a PS fuse that WASN"T SOLDERED!
It was tightly twisted on the terminal though, and made connection for decades.
DAH!
Now on to the last one...I said plastic tiger, but actually it's the 215...a two channel lower power version of the Tiger.01.
Les
you buy the beer and i'll talk microphones till i'm blue in the face!
The other Tiger .01 is fixed. It was just a bad connection. Hate to admit it, but I found a connection to a PS fuse that WASN"T SOLDERED!
It was tightly twisted on the terminal though, and made connection for decades.
DAH!
Now on to the last one...I said plastic tiger, but actually it's the 215...a two channel lower power version of the Tiger.01.
Les
Last edited:
I Buy, You Fly...
'Mere to Lebnun, Tn. near the first of a month, before I've
spent my disability pension, with a vehicle for us to get out
'n' get us a beer in, 'n' you got yourself a deal!
'Mere to Lebnun, Tn. near the first of a month, before I've
spent my disability pension, with a vehicle for us to get out
'n' get us a beer in, 'n' you got yourself a deal!
SWTPC Universal Tiger MKII schematic
The Universal Tiger MKII used the same output devices as the original article version. They were the MJ802 and MJ4502. It just had two extra transistors that were for protection. I need the schematic of the MKII version if anyone has it available.
Thanks,
Phil Cline
The Universal Tiger MKII used the same output devices as the original article version. They were the MJ802 and MJ4502. It just had two extra transistors that were for protection. I need the schematic of the MKII version if anyone has it available.
Thanks,
Phil Cline
Found the bug in the SWTPC designs, finally!
http://www.diyaudio.com/forums/soli...-tiger-improved-simulation-2.html#post2585028
I wrote there:
"OK, well I think I've found the fundamental bug in the Universal Tiger (UT). When I run the sim and probe the Vbe of the drivers (Q5, and Q6) I saw a reverse Vbe of 7.5 volts without clipping, and about 18 volts in clipping. The max rating for most transistors is about 5 to 10V, I did not see a spec for the original devices used in the UT but the MJE243 for example is rated for a maximum of 7V. Exposure to this voltage will damage the B-E junction and with enough reverse current burn it out."
http://www.diyaudio.com/forums/soli...-tiger-improved-simulation-2.html#post2585028
I wrote there:
"OK, well I think I've found the fundamental bug in the Universal Tiger (UT). When I run the sim and probe the Vbe of the drivers (Q5, and Q6) I saw a reverse Vbe of 7.5 volts without clipping, and about 18 volts in clipping. The max rating for most transistors is about 5 to 10V, I did not see a spec for the original devices used in the UT but the MJE243 for example is rated for a maximum of 7V. Exposure to this voltage will damage the B-E junction and with enough reverse current burn it out."
I've used the B-E junction of small signal transistors as a zener diode with no problem. It doesn't seem to hurt anything at low currents. The question is how much the Tiger devices can stand vs how much they get subjected to.
There was the original MJ802 and 4502, but I know they subbed a TI part (as did I). Is that one of the ones you simulated? I always thought the amp was slightly "different" when the TI part was used.
Tigers are incredibly sensitive to layout and construction. I don't know if the original article version worked by luck or design, but I built some real disasters until I figured out what it needed for power distribution and routing. IMO, there's a huge amount going on that won't make it to the simulation; I'm not sure what or where you'd add it if you wanted to.
There was the original MJ802 and 4502, but I know they subbed a TI part (as did I). Is that one of the ones you simulated? I always thought the amp was slightly "different" when the TI part was used.
Tigers are incredibly sensitive to layout and construction. I don't know if the original article version worked by luck or design, but I built some real disasters until I figured out what it needed for power distribution and routing. IMO, there's a huge amount going on that won't make it to the simulation; I'm not sure what or where you'd add it if you wanted to.
I had previously simulated with the OnSemi models for the MJ802/4502 but I do not trust those models.
While you can use a reverse biased junction as a Zener, even a diode, you would always use current limiting, and I'm not sure where the current limit is for the amp - something worth looking into. BJTs are not generally designed to be used in this way:
IEEE Xplore - Effect of emitter-base reverse bias stress on high frequency parameters of bipolar transistors
I believe that what you refer to as "a huge amount going on" is probably that the CFP output is close to instability making it sensitive to layout and power distribution.
I don't know why you say it won't make it to simulation, simulation works very well with good models.
While you can use a reverse biased junction as a Zener, even a diode, you would always use current limiting, and I'm not sure where the current limit is for the amp - something worth looking into. BJTs are not generally designed to be used in this way:
IEEE Xplore - Effect of emitter-base reverse bias stress on high frequency parameters of bipolar transistors
I believe that what you refer to as "a huge amount going on" is probably that the CFP output is close to instability making it sensitive to layout and power distribution.
I don't know why you say it won't make it to simulation, simulation works very well with good models.
Last edited:
What are better spec transistors for Q5 and Q6? I don't recall what I have in my B's. Since this thread won't die, I mean keeps getting better, we should come up with a list of replacement parts to keep these amps alive.
I'll start:
Bigger PS caps. I'm using 10,000uF 75v Sprages and haven't had a PS failure since the upgrade. They do not fit in the original chassis. If someone found similar caps (Mouser source?) that will fit in the chassis, please post.
I actually used MJE243/253 for Q5 and Q6 years ago and suggested them here, however after looking closely at the SOA for these parts I have to say that I would not use them again. They do not see the rail to rail voltage in this design and therefore an 80V (Vce) device such as the BD139/140-16 should work fine and these have significantly better SOA. Use a heatsink with them. But again, I really can't suggest this design until the issues are worked out.
What I was referring to above were strays. Unless you put in lead resistances and inductances, and probably even small capacitances to the chassis, the simulation doesn't reflect the real circuit. It certainly shows you where the weaknesses are, but I found that seemingly minor construction differences, where HF bypassing and the Zobel network were connected for example, made the difference between decent stability and constant oscillation. The value of the Zobel is also critical. The Tiger seems way more sensitive than other amps I've built, yet with the right construction it's reliable and doesn't oscillate. I assume it comes down to phase margin- make some unfortunate decisions in the physical construction and all bets are off. I built a very pretty one in a small rack chassis and it was junk. Nothing I did would cure the slight oscillation on musical peaks. I rebuilt it with the exact same components in a different physical configuration and it was stable as a stone.
Hello everyone,
I have been following this thread for a while and there is an awesome amount of information on how to make the UT better. I want to thank everyone for the information given. I have a couple Universal Tiger MKII (with the extra SS transistors) but one of the boards is damaged. I was wondering if anyone had an etching guide like the one that came with the original UT. I have looked everywhere including the SWTPC archive site and can only find the etching guide for the original UT.
Thanks,
Bobby
I have been following this thread for a while and there is an awesome amount of information on how to make the UT better. I want to thank everyone for the information given. I have a couple Universal Tiger MKII (with the extra SS transistors) but one of the boards is damaged. I was wondering if anyone had an etching guide like the one that came with the original UT. I have looked everywhere including the SWTPC archive site and can only find the etching guide for the original UT.
Thanks,
Bobby