I built a pair of Lateral Mosfet Amps (Renesas LMosfet Randy Slone designs from his High Power Amp book) and they both ended up with >100mv of DC offset at the speaker outputs. Recently I beta matched a bunch of small signal transistors for the differential input pairs and soldered them into the amps. This fixed one amp, I was able to rebalance the differential input with the balance trimpot and rebias to 45 mv idle and had < 15 mv DC offset. The other amp remained at >100 mv DC offset, and has some audible hum from the speaker. Is this a cap problem? Or do I need to try different differential input pair transistors? I have a sig gen and an oscope. Can these be used to track down the source of the DC offset? I can post a schematic for the amp tomorrow if it is helpful.
Thnks!
Thnks!
Hum will probably be down to wiring problems, ground loops etc. Input and feedback ground need to be referred to a common "star point". If you are using a connection taken between the main reservoir caps then that's not good enough.
DC offset is down to the circuit design... 100mv is a lot. Can't really comment without knowing the circuit.
DC offset is down to the circuit design... 100mv is a lot. Can't really comment without knowing the circuit.
Yes, both amps are grounded identically. One is the right channel, one is for the left. I guess I'll have to pull the board and go over it with a magnifying glass and check for cracks or poor solder joints.
It looks like I should add a schematic of the design I built. It might make it easier to figure out some possible problems areas for DC offset.
On this particular design, the differential input pairs apparently need to be closely matched (and there are a pair of pairs to add extra complication, a pair of npn, and a pair of pnp transistors). Any mismatch presumably can be tuned to a minimum via a trimpot... it balances the two pairs ). I wish I could remember the name and exact functionality of the scheme. Anyway it worked out well on one amp with < 15 mv of DC offset, but not so well on the other. However, I'm not sure if this is the only area of possible mismatch in the system. I need to break out the oscope and compare the two channels to see what differences I can detect. I need to get another set of scope leads I suppose. It will be easier to compare the two amps if I do. Will the signal with DC offset appear to be riding on DC? Is there anyway to see the DC offset with an Oscope? I look at trigger pulses and waveforms on oscopes all the time for my work. This, however, falls outside my experience.
There seems to be a certain ammount of poor regard for Mr. Slone's amp designs. I can say that some simpler designs I've built have worked out better... especially for the novice DIYer that I am.
It looks like I should add a schematic of the design I built. It might make it easier to figure out some possible problems areas for DC offset.
On this particular design, the differential input pairs apparently need to be closely matched (and there are a pair of pairs to add extra complication, a pair of npn, and a pair of pnp transistors). Any mismatch presumably can be tuned to a minimum via a trimpot... it balances the two pairs ). I wish I could remember the name and exact functionality of the scheme. Anyway it worked out well on one amp with < 15 mv of DC offset, but not so well on the other. However, I'm not sure if this is the only area of possible mismatch in the system. I need to break out the oscope and compare the two channels to see what differences I can detect. I need to get another set of scope leads I suppose. It will be easier to compare the two amps if I do. Will the signal with DC offset appear to be riding on DC? Is there anyway to see the DC offset with an Oscope? I look at trigger pulses and waveforms on oscopes all the time for my work. This, however, falls outside my experience.
There seems to be a certain ammount of poor regard for Mr. Slone's amp designs. I can say that some simpler designs I've built have worked out better... especially for the novice DIYer that I am.
One of his designs is not workable, something tells me this is what you tried to build. This design suffers from high and unstable dc offset. Could you show us a schematic.
Slones designs arent bad, theres problems with one of them, the others work perfectly, then there is his amps he use to sell called optimos. Well now, build yourself a pair and youll have a very big surprise how good these amps sound and measurements are impressive too. Most underrated amp around this forum I would think, nothing special to its design except one particular stage, Ill say no more, theres one thread relating to the sound quality of these amps that says it all.
Javajiven if you have built the design im thinking you did, you wasting your time youll always have problems with it, youll have to start over from scratch, I can supply you the optimos schematics privately if you interested in building it. Parts are about the same and interchangeable.
Rest in Peace Mr Slone.
Slones designs arent bad, theres problems with one of them, the others work perfectly, then there is his amps he use to sell called optimos. Well now, build yourself a pair and youll have a very big surprise how good these amps sound and measurements are impressive too. Most underrated amp around this forum I would think, nothing special to its design except one particular stage, Ill say no more, theres one thread relating to the sound quality of these amps that says it all.
Javajiven if you have built the design im thinking you did, you wasting your time youll always have problems with it, youll have to start over from scratch, I can supply you the optimos schematics privately if you interested in building it. Parts are about the same and interchangeable.
Rest in Peace Mr Slone.
Rip
that's too bad ... i've talked to him; he was a really nice guy, knew quite a bit and liked talking audio.
mlloyd1
that's too bad ... i've talked to him; he was a really nice guy, knew quite a bit and liked talking audio.
mlloyd1
virtually every modern ( last 40 years or so) oscilloscope's input amplifier is dc coupled. Somewhere near the input connector is a switch labeled AC DC. The problem you may have if you are using a 10x probe is that if the scopes sensitivity is to low you may not see 10 mv of offset very well. Use a 1x probe or see if there is a switch on your probe that says 1x 10x. 10 x means 10 times higher input impedance at the cost of 1/10th vertical sensitivity. If you scope has 10mv/div sensitivity and your probe is a 10x probe and you are measuring a 10mv signal the display will move about the thickness of the horizontal trace. If your probe is 1x or switchable the trace will move up or down 1 division up or down depending on +/- input. The display will probably be a bit fuzzy as well at those millivolt levels. With a 1x probe the input resistance of your scope will more than likely be 1 meg ohm.
When taking any measurements with a 1x probe you are putting a 1 meg resistor in parallel with whatever circuit you are testing to ground. For SS stuff this will have minimal effect in most cases, with tubes and their high value grid resistors, plate resistors etc a 1x probe will change the circuit parameters! That's 1 reason why 10x probes and higher exist.
When taking any measurements with a 1x probe you are putting a 1 meg resistor in parallel with whatever circuit you are testing to ground. For SS stuff this will have minimal effect in most cases, with tubes and their high value grid resistors, plate resistors etc a 1x probe will change the circuit parameters! That's 1 reason why 10x probes and higher exist.
Yes, I just saw the news that he passed away. Now I really need to finish the amp as a tribute to the Man. I must say I do feel guilty for my comment concerning the regard for his amp designs.
Homemodder: I was able to get one of the amps adjusted down to below 15 MV of DC offset. Hopefully I'll be able to accomplish the same for the second. I'd post a schematic except I've been very busy at work and don't have a scanner at home. If I can't get the DC offset problem adjusted down to an acceptable level, I'd definitely be interested in the Optimos designs.
I found some old posts where they gave someone advice concerning DC offset troubleshooting in Randy's amp. Using this info, I should be able to figure out if the problem is in the input differential and current mirror stage this weekend.
Homemodder: I was able to get one of the amps adjusted down to below 15 MV of DC offset. Hopefully I'll be able to accomplish the same for the second. I'd post a schematic except I've been very busy at work and don't have a scanner at home. If I can't get the DC offset problem adjusted down to an acceptable level, I'd definitely be interested in the Optimos designs.
I found some old posts where they gave someone advice concerning DC offset troubleshooting in Randy's amp. Using this info, I should be able to figure out if the problem is in the input differential and current mirror stage this weekend.
Javajiven, thats exactly the problem, you cannot use dual differencials with current mirrors, if you look carefully at the design youll see that with a arangement like this the vas current isnt defined, even if you were to get lucky to get offset low because of device tollerances like you maybe did in your one channel you see that the offset will vary with volume, you could get high dc with increase in volume and your bias for the output devices will vary according to volume too creating eiher crossover distortion or worse, Im afraid youll have to scrap that design. His optimos is practically the same design without this specific problem.
If you do a search here there is a thread about this design problem, no solutions were forthcoming except if you change the design completly.
Be aware you could have low offset with no input but it could rise unexpectedly with input to levels where it could do damage to your speakers.
If you do a search here there is a thread about this design problem, no solutions were forthcoming except if you change the design completly.
Be aware you could have low offset with no input but it could rise unexpectedly with input to levels where it could do damage to your speakers.
Homemodder: I connected the amp with low dc offset to a cheap speaker and cranked it up. You are correct about the design. The offset ranged from 10 mv to about 700 mv... Well, I guess not all his designs were a great success, bless his soul.
Thanks to everyone who posted answers to various aspects of my post, esp Multisync for your insights into Oscope function.
Thanks to everyone who posted answers to various aspects of my post, esp Multisync for your insights into Oscope function.
Dont worry I built this design a lot of years back too and had to scrap it, also learned the hard way and built it without carefully studying the circuit first and missing the design flaw.
Since you have the parts already I would recommend you now try his optimos circuit, you wont have problems here and have a very good amp amp. It only uses two additional small signal transistors as cascodes for the vas which are referenced to the output which is the strong point of this design and turns those current mirrors into cascodes too. The vas cascodes kill high frequency distortion very effectively and its probably this that gives it very good sound quality from a relatively simple circuit. If you read the other threads about the optimos youll notice owners compare this amp favourably to designs which cost 20 times more and names like Ayre come up. Its one of the better sounding amps I have in my home too. Another exceptional design and highly underated amp is the stochino design, but its much more complicated and difficult to build. On my list of my top 3 good sounding class AB amps from this forum.
Yes probably a lot of people tried to build the design. It was one of only 2 Mosfet designs he provided circuit board trace templates for. It seemed the better of the two designs that I could easily etch PC boards for (without having to draft the PC Board layout myself) so I went for it. Im not really savy enough to catch the design limitations Homemodder pointed out. Anyway, it ended up being a challenging educational experience.
The design I'm referring to is from Randy's later book on Amps, "The Audiophile's Project Sourcebook: 80 High-Performance Audio Electronics Projects". Its known as Design Fig. 6.21 in that book. Appearently it became know as the Optimos at some point. I've managed to track down the schematic to some website that archived the Randy's defunct Seal Electronics webpage. http://web.archive.org/web/20040612124552/www.sealelectronics.com/kits/images/OPTI1001.pdf is a link to the schematic.
I was curious if q1-q4 need to be matched.
I was curious if q1-q4 need to be matched.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.