Advice on setting bias... |
Post Reply | Page <1234> |
Author | |||
M4trix
Registered User Joined: 23 June 2009 Location: Croatia Status: Online Points: 445 |
Post Options
Thanks(0)
|
||
Before I get attempted, what are the voltages on the power supply ? +/-50-60VDC ? Edit: OK, it's +/- 52VDC according to JBL's document. Edited by M4trix - 03 August 2020 at 6:22pm |
|||
monkeypuzzle
Young Croc Joined: 25 January 2005 Status: Offline Points: 923 |
Post Options
Thanks(0)
|
||
Cheers, I can't remember what the voltage was, I want to fix this before I fire it up again just incase something fries.. Its just really odd that one way it reads half what it should and the other way it reads so incredibly high, almost like the resistor is acting like a diode... This is measuring directly off the solder that surrounds the resistor leads that come through the board so theres no track in the circuit... They are good joints as well. I'll get the glass on it and look over the tracks in details later though and give the T03s another check over as you suggest, cheers.
|
|||
blah blah blah blah blah......
|
|||
M4trix
Registered User Joined: 23 June 2009 Location: Croatia Status: Online Points: 445 |
Post Options
Thanks(0)
|
||
I'm almost there. Some BIAS problems... Edit:
Well, MD8003 (dual transistor) is dead as a Dodo bird and 2N3440 as
well. No simulation model for those so I fitted two 'low noise' 2SC2240
instead. Still have problems with the quiescent current which is waaay too
high. I'm pushing the BIAS pot wiper all the way down but to no avail.
Heck if I knew what is in the output. Tried MJ15003 and 2N3773. Anyways,
this is still a simulation which is far from ideal. P.S. Well I'm doin' my best ! Edit 2: Ok, bypassed R17, changed R30 & R32 to 47R, replaced MJ15003 with 2N3773 and the quiescent current is now 50mA which is acceptable. Probably MJ15003 has a higher gain. The BIAS wiper is still in the corner. LOL Edit 3: I'm done. Changed R30 & R32 to 22R. BIAS trimpot is now useful and the quiescent current can go down to zero. LOL. Testing under full load with some slight clipping --> Any suggestions guys ? Edited by M4trix - 04 August 2020 at 4:18am |
|||
jbl_man
Moderator Group Joined: 12 January 2005 Location: London. Status: Offline Points: 11154 |
Post Options
Thanks(0)
|
||
Well done,and a schemematic my eyes can actually see too!
|
|||
Be seeing you.
|
|||
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
||
I like those mixed signal Agilent scopes, same unit I still have in the garage!
Looking at the circuit schematic quickly it seems to me: R30 and R32 should be sized to bias Q13 and Q14 on depending on the collector-emitter current flow of Q11/Q12 Q11 and Q12 are common emitter, so should be fairly independent of beta for voltage gain. It's a slightly weird circuit, you've almost got 2 sources of bias. I have to say I don't quite understand the point of the R20/R21 voltage divider. On first glance you'd think that was setting a voltage that D6 added to and then set the base of Q7. But that can't be the real mechanism going on the resistor values and then you look again and the voltage between R20 and R21 must almost be equal to the ampifier output looking at the Vbe and diode drop in parallel. Anyway, that's a bit odd I feel. It seems like Q7 and Q8 exist to define a relationship between voltage rails/ground and the amplifier output in some way. But its hard to see what drives what. As you've probably figured, the parallel combo of R14//R15 is the bias adjustment and is paralleled to approximate a logarithmic adjustment. Q6 is the Vas stage, Q4/A5 a constant current source loading it. R17 sets a minimum bias which adds to the voltage drop over R14//R15. Bigger voltage here over all resistors = more bias. You're also missing the equivalent of R161 feeding the negative rail output transistors. Is that right? |
|||
M4trix
Registered User Joined: 23 June 2009 Location: Croatia Status: Online Points: 445 |
Post Options
Thanks(0)
|
||
Hi odc04r, well the original schematic contains serious errors. Actually it's not an original schematic. The guy traced the tracks from the PCB and made some flaws/mistakes. The original source is here -->
Neither do I. They are making problems in the simulator as well. The amp is working fine even without them so I grounded the suckers. Maybe they're bleeding resistors !? LOL
Yes, you're right. They are R13 and R18 now. Thanks for the tip ! Also, R23 & R24 had wrong values (1k5). The Q7/Q8 limiters weren't kicking in when short circuit was applied at the output. The proper value is 150R. The limiters (when short circuit is applied) now keeps the current around 3.1A per output transistor which is, according to the 2N3773 datasheet, still in the SOA window. That's all folks. Btw, where is monkeypuzzle ? Yo mate ! Here is the proper schematic... well...I think. Still waiting for suggestions. Cheers. |
|||
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
||
Looking better! Yes a weird voltage divider that, one of those things perhaps that worked on a bench but the designer wasn't sure why and stuck it in anyway :) Maybe it affects the attack of the clipping circuitry in some weird non-linear way. I'm not quite sure what D6/R21 are doing either. They are acting as a second limiter to the limiting circuit by making sure the base of Q7 can't go higher than the D6s forward voltage and then R15 limits the current through it in that case. Maybe they are a safeguard to avoid blowing up Q7 with excess base current. Q7 is going to conduct when the voltage over the output 0.15 resistor is driven up to ~0.6V by output current. The voltage drop over R23 will be minimal, so that exists just to set the base current which will then define how hard Q7 turns on once the voltage trip is reached. When Q7 comes on it steals base current from Q9, which then reduces the output of the amplifier. So I don't really see why you need D6/R21 assuming that R23 was correctly sized in the first place. I would definitely spend a while simulating the clipping behaviour of that circuit! The voltage gain driver followed by compound pair is a little odd too. I guess the TIPs are just there to get around the quasi-symmetry of the output transistors. These days if the 2n3773Gs have a p-type equivalent you could probably design out the TIPs, and maybe uprate the 2n3439s to cope driving direct if necessary. Depends on how you feel about historical accuracy as always.
Edited by odc04r - 05 August 2020 at 9:28am |
|||
monkeypuzzle
Young Croc Joined: 25 January 2005 Status: Offline Points: 923 |
Post Options
Thanks(0)
|
||
Okay, so maybe this is a little clearer. I've taken out the confusing bits regarding the 500 series so we're just looking at 300/302. A few things to note, The TIP29/30 may not be what's in there. They are also Turner branded as Turner 1361 and 1371. The one repaired on the russeltech site has a TIP29 in it so at some point it's been replaced in its history and this has worked. The board he repairs is marked A2PM 1062, all mine are A2PM 1064 and this diagram is taken from them. On my drawing the component identifiers are from the actual board to help me with fault finding but for the sake of this thread I'll go off the plan supplied by M4trix as thats where everyones head is at... Over the range I have there are some that have bypassed R17 as you've found, there are some that have R10 piggybacked with either a 1k2 or 2k7 giving 337r and 400r respectively. There isn't a pair to R161 and as for output devices, I asked robert at russeltech and he recalls using MJ15024. I found the problem with the weird readings I was getting the other day... operator error.. I still had my sig gen plugged in which was mucking up my meter even though the power was off, the board was out of the socket and only have the audio input wired in..... doh I didn't manage to bring back a working amp yesterday but will do that probably later today and then I'm going to try and get the broken board in front of me going and I'll use that to test it out with some actual TIPs in it and I also have a pair MJ15024 devices which I'll drop in as well and take some measurements. Please bear with me though, My understanding fades out somewhere between A level and the RA Penfold books of my youth however I am enjoying this voyage of learning and discovery! All your comments and interest is greatly appreciated. However these work, they may not chuck out loads relative to modern or even other quite old options but I've ended up with quite a lot of them for not very much money and they do sound very nice..
|
|||
blah blah blah blah blah......
|
|||
M4trix
Registered User Joined: 23 June 2009 Location: Croatia Status: Online Points: 445 |
Post Options
Thanks(0)
|
||
Ok, I ditched TIP29/30. They aren't real complementary pairs anyway. So I replaced them with MJE15030/MJE15031 and the outputs with MJ15003 ( MJ15024 is an over-kill for this task ). With those changes I have so much better results ! Btw, 2N5415/2N3439 aren't complementary pairs either so I have to ditch them as well. Complete results later. Need some rest. Well I also have a life outside Speakerplans ! Cheers. Edit: @monkeypuzzle, so you found the error. Good.
Edited by M4trix - 05 August 2020 at 6:35pm |
|||
jbl_man
Moderator Group Joined: 12 January 2005 Location: London. Status: Offline Points: 11154 |
Post Options
Thanks(0)
|
||
MJ15003 is probably more in-keeping with the vintage feel of the amp, although the MJ15024 are a superior rating,they of course,weren't around at the time this amp was made.
I have a pile of original RCA MJ15003 pulled from various vintage amps of the mid/late 70's. |
|||
Be seeing you.
|
|||
odc04r
Old Croc Joined: 12 July 2006 Location: Sarfampton Status: Offline Points: 5483 |
Post Options
Thanks(0)
|
||
I'd say you're doing pretty well! Very satisfying bringing a dead amp back to life. I am about to rebuild one of my old earlier creations in the next weeks. Every few years I rebuild and improve a part of it. This time it is redesigned PCBs and intend to then re-build the PSU PCB with the option for active regulation. Which is quite pointless really, but everything is a learning experience if you want it Lastly I am going to build in some true RMS power monitoring of outputs via microcontroller and then maybe also mute outputs under fault conditions. Could also monitor bias and heatsink temps. It's easy to get carried away... |
|||
monkeypuzzle
Young Croc Joined: 25 January 2005 Status: Offline Points: 923 |
Post Options
Thanks(0)
|
||
I've got two A500s back online again as they were apparently plugged into each other via a pin swap speakon cable and this popped out part of the earth track on the PCB. Everything else on the boards survived however and a track repair has them going! Another one lost one of the TO220 devices with R30 burning out with it. Again, everything else metered out fine so a swap out from a donor unit and it's up and running again. On the one on my desk there's +44.7v on the output. All my tools are currently at the workshop at the moment apart from my really rubbish emergency purchase multimeter so I can't do a whole load of testing at the mo but if anyone has any thoughts or suggestions of how to go about properly fault finding this as I'm sometimes stumbling around in the dark that'd be great. Here's a very quick pic from a few weeks ago, the lower fans blow in and the upper push out, they get warm and I just don't want to risk it. I have two of these stacks and thats as far as it's going to go. 2 A500s (single TX but twin PSUs) driving bass, one B502 doing mid and a B302 on the comps. Warm white LED tape routed in to the wood just at the top... need to put a dimmer on it! Will take some pics of the back when I next go in. |
|||
blah blah blah blah blah......
|
|||
Post Reply | Page <1234> |
Tweet |
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |