Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 5th June 2006, 06:48 AM   #11
Shredly is offline Shredly  United States
diyAudio Member
 
Join Date: May 2006
With 100mA of quiescent (shorted input) current at the emitters of the MJ15003/004s, the amp sounds just a bit better than it did with the original MJ15001/002s; the biasing of the Class-A voltage amplifier was not significantly disturbed; it is still within a very small percentage of the center of the load line, in fact it is just slightly better than it was. The power supply regulation project failed, though; I will now buy a toroid with two secondaries and make it work. Anybody wanna buy a used transformer, cheap? Just kidding. It's bench supply meat. I may build some more Lin amps and it would make a pretty good test supply for that.

Direct-coupling the heatsink does seem to have improved the hum slightly, but that may be subjective, and I intend to be picky. I've also observed that there is significant hum in the preamp as well, so it's clear to me that regulated power is mandatory in this application. That was there all along, and of course it doesn't matter for that what I do to the Lin amp. I decided to leave ground alone other than to ground the heatsink; after I get the supply regulated, I'll run an A/B taste test to see whether a capacitor or jumper is more appropriate, I couldn't tell for sure with all the hum from the ripple. :P

Hey, I tried bypassing the base of the diff amp current source transistor to ground and then to +40V with a 0.1uF cap, and it didn't do a dang thing for me. I thought of increasing the 220uF cap to ground from the +40V rail right next to where the Zener gets its supply, but didn't bother doing anything about it; my expectation is that regulating the power supply will correct the problem if that's it.
  Reply With Quote
Old 16th June 2006, 07:15 AM   #12
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Hi Shredly,
is it hum and it's harmonics that's the noise problem or wide band noise (white or pink)?

There's a mistake in the schematic;- C16 is shown connected to top end of R28. It should be bottom end of R28 to C16.
Since R28 is 0r33 it may not make a lot of difference.

The protection circuit seems to be a little unusual and I suspect not very efficient, particularly in view of the mishap that caused the overheating.

I note you managed to bias the output stage but your description of the procedure you were planning to follow could have blown up the output stage. The label saying current mirror is actually next to the Vbe multiplier (as poster said). You seem to understand the effect of changing resistor values but for first start up after component changes you should firstly set the multplier to MAXIMUM resistance to MINIMISE the voltage bias across the output stage. Secondly you would be wise to ALWAYS start up a project on test through a mains light bulb to reduce the risk of blow up.

However it worked and you got away with it.

Your Email is off.
__________________
regards Andrew T.
  Reply With Quote
Old 16th June 2006, 07:17 AM   #13
Did it Himself
diyAudio Member
 
richie00boy's Avatar
 
Join Date: Nov 2003
Location: Gloucestershire, England, UK
Just to add to Andrew's post, if it's hum it's likely a ground issue.

There is another fatal error in the schematic, that is that the right leg of the LTP is connected not to the same -ve rail as the rest of the circuit, but to the VAS emitter, which is incorrect.

The LTP CCS is also wrongly implemented. Either R4 or D3 will serve no purpose and the Vref generation I suspect with just a quick look is all wrong. It's also bad practice to take out the feed from Vref like that as you open it up to picking up noise etc.
__________________
www.readresearch.co.uk my website for UK diy audio people - designs, PCBs, kits and more.
  Reply With Quote
Old 16th June 2006, 07:46 AM   #14
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Hi Richie,
my circuit analysis is not too hot but I think the VAS is connected as a folded cascode. It looks similar and the collector resistor ratio looks about right for that topology. R9 voltage drives the base in common emitter (inverting) and R11 voltage drives the emitter in common base (non inverting). Output from VAS is inverted from both sides of the LTP.

The CCS generates a constant voltage across the series pair of R4 and Vbe of Q3. This forces a constant current ([10-0.7]/2.7=3.4mA) through the resistor and C10 tries to hold the constant voltage steady when dynamically loaded.
__________________
regards Andrew T.
  Reply With Quote
Old 16th June 2006, 09:29 AM   #15
Did it Himself
diyAudio Member
 
richie00boy's Avatar
 
Join Date: Nov 2003
Location: Gloucestershire, England, UK
Hi Andrew. As I said I only quickly looked at the circuit. Now I see that the CCS is OK - what threw me was R8 which would either screw up the CCS totally or (in this case because it's only 22R which I didn't notice before) just cause slightly poorer CCS action.

You could be right about folded cascode, equally it could be a drawing error.
__________________
www.readresearch.co.uk my website for UK diy audio people - designs, PCBs, kits and more.
  Reply With Quote
Old 16th June 2006, 02:59 PM   #16
Shredly is offline Shredly  United States
diyAudio Member
 
Join Date: May 2006
Thanks, you two. I appreciate the feedback. I'll double-check the agreement of the schematic with the one I dug up from the archives; I've already checked that drawing against the circuit board, I'll just make sure I didn't screw it up while I was transferring it to my CAD software.

Bear in mind that this is a commercial product, produced in the late 1980s. It's not my design, and the only two changes I made were to replace the output Darlington pairs and to replace a cap that was decoupling the heatsink to ground with a jumper, after ensuring that the heatsink wasn't connected to the transistors.

I found some instructions for biasing a very similar Lin topology amp at AmpsLab; after looking them over and making sure I understood what they were doing, and looking at my schematic, it was clear to me how to proceed, and that was not quite the instructions I posted here. I was actually rather put out by the fact that I never got any response to my request for review of my procedure, and went looking for myself; I wasn't going to wait around forever.

Since hFE was not sufficiently different to introduce a great deal of bias change at the bases, particularly through a Darlington pair (which is one of the advantages of the Darlington pair), I decided to begin by leaving the settings of the two variable resistors alone; because I was careful during disassembly, I was confident they were set where they had come from the factory. What I did was connect to the bases of the transistors, set my meter for sample-n-hold, and flip the power switch on and off quickly. The values looked pretty good. Next I did much the same across the current limiters, and finally across the A-class amplifier (the one that's connected at its collector to the stuff I labeled "Current Mirror." I chose that name, by the way, because it helped make it clear in my mind what it was doing, not necessarily because that terminology was technically correct- I am aware that a real current mirror is at the collectors of a diff amp pair).

Everything looked OK, so I flipped the power on for longer and started checking for overheating with a thermocouple probe. Again, everything looked OK, so I checked the bias by checking the voltage across the resistors between the output MJ15003/04 pair; 100mA was the quiescent current value suggested that should put them up above the knee and into the active region, and their datasheets confirmed this, so that's what I set it for. The adjustment was very slight on the PNP (bottom) side; and far less than a tenth of a turn on the NPN side as well, indicating that in fact the Darlington pairs were functioning correctly and compensating for the hFE and consequent DC Beta differences.

I'll get back to you on the accuracy of the schematics.
  Reply With Quote
Old 16th June 2006, 03:11 PM   #17
Shredly is offline Shredly  United States
diyAudio Member
 
Join Date: May 2006
Oh, and most important:

1. It's 120Hz hum; I'm not hearing much in the way of harmonics, just straight 120Hz, fairly standard for a bridge rectifier-filter cap setup. The amp is pretty clean, I don't hear much harmonic distortion (unless I add some in ) , and there's very little white or pink noise; I generally run pre-out of this amp to a very nice clean board, so I can put it through some signal processing hardware I use, then feed it back to the power in, and I record off that board as well. There is a fair bit of pink noise from the preamp, but that's not unusual, and they've used bifet opamps and a few other tricks to cut it to levels that were unusually low for that era. Running a clean signal into the power amp, I get pretty much what I expect, with the exception of that 60Hz hum, and it's that hum that I object to. And it's not something that started with the new transistors; it's been there all along, since I bought the amp, and I bought it brand new off the music store floor.

Bear in mind that the network at lower right appears to be modifying the tone; this is not intended as the power amp for a high-fidelity system, they're trying to replicate some of the tonal characteristics of a 6L6-based or similar tube power amp output. I've been thinking I might check for an attempt on the part of the original designers to put a 120Hz filter in there, too; I don't think the values are right, the rolloffs don't quite look right for that to me, but I'll double check. My personal opinion is that it's a better idea to prevent it in the first place, at the input before you start amplifying it, but these guys did some other pretty weird stuff too; for example, the whole story of the, honestly, folks, THREE freakin "grounds" running right next to each other, I still can't believe that crap. Somebody bought into the mythology there, big-time.
  Reply With Quote
Old 16th June 2006, 03:23 PM   #18
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Hi,
there is absolutely nothing wrong with running multiple grounds to an audio Central Star Ground. For the moment I shall assume the designer got his multiple grounds right.

What is much more likely is that the safety earth and the audio ground have been connected. This is probably the source of your hum.

The finger pointing at Zeners etc may have accounted for wideband noise, but you have now discounted them.

There is a big problem here.
Guitarists get killed if the grounding is unsafe.

You need to be absolutely sure of the safety side of your grounding and how it interacts with your metal guitar strings BEFORE you contemplate ANY changes to your pre or power amp.

I could advise in a non professional way but at either end of a forum I cannot guarantee some misunderstanding will not lead to a wiring error that could be unsafe.
So, for your safety I bow out.

Your Email is still off.
__________________
regards Andrew T.
  Reply With Quote
Old 16th June 2006, 03:39 PM   #19
Shredly is offline Shredly  United States
diyAudio Member
 
Join Date: May 2006
No, I absolutely confirm C16 is present both on the original design schematics and on board. That's precisely how it's built. It connects the emitter of Q10 to the voltage divider of R15 and R16.

I also confirm the existence and accuracy of R8, R4, D3, and R18, as well as R9 and R11. There are no drawing mistakes. The schematic you see is the circuit board in my amp.

My impression of C10, BTW, was that it acted with R12, C6, and R10 to cause a 1-pole rolloff above 4.8MHz and below 0.26Hz, for DC unity gain and high-frequency stability, promoted by the negative feedback to the diff amp Q1 and Q2.
  Reply With Quote
Old 16th June 2006, 04:11 PM   #20
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Hi Shredly,
the CCS and input filters are all OK.
The NFB DC block and bass roll off are set excessively low but should not cause a problem. Slight performance benefits may accrue from adjustment to the input filter (59Hz) and NFB roll-off (22Hz) frequencies, but due to the specialised nature of the amp may not be worth the effort to alter.

The main high frequency roll of is due to C8, the Miller compensation cap. This is a cheap and cheerful method of ensuring stability in the amplifier and a little bit of stability adjustment from C10 is usually all that is necessary for a well behaved amp in mass production. It works and it is often used.
Prior to the C8 roll off, the input filter, as you have correctly noted, reduces the audio response to -3db @ 15kHz, about -1db @ 7.6kHz. In view of the harmonics the guitar is capable of and your speaker response do you think there is any advantage in opening up the HF filter by say half an octave?

Do you use the mix out or the pre out? They both have medium to high source impedance. Very short cables and low input capacitance on the next stage are vital.
__________________
regards Andrew T.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Rebuild of old box and driver Moondog55 Multi-Way 1 25th May 2009 04:00 AM
Some topology of current feedback amplifier aht Solid State 7 13th August 2005 02:49 PM
Hawksford wrong on amp topology for current driving loudspeakers? capslock Solid State 5 7th November 2003 02:43 PM


New To Site? Need Help?

All times are GMT. The time now is 07:41 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2