I already had it powered with output stage and working OK only at lower voltage (25vdc). So no errors in assembly to look for. It was higher voltage that blew outputs.
I am ordering a new pair of (hopefully now original) MJE15032/MJE15033 to verify that the fake drives were the culprit.
I also noticed that the variable offset resistor can give spikes of offset current when used with amp running. From now on I will be making offset changes only in powered off state.
I am ordering a new pair of (hopefully now original) MJE15032/MJE15033 to verify that the fake drives were the culprit.
I also noticed that the variable offset resistor can give spikes of offset current when used with amp running. From now on I will be making offset changes only in powered off state.
I did that last week.
I installed everything up to the drivers including the driver emitter resistor.
I added in two 1k0 resistor from the respective driver emitters to the output line. These two resistors feed a split signal back through the NFB route to the -IN node.
I expected that to allow the amp to settle.
It did not work as I expected.
In the end I had to bite the bullet and install one pair of outputs, turn down the Vbe multiplier (set VR to maxium resistance) and power up via a current limited lab supply.
That worked, even though I had forgotten to remove the two temporary 1k0 resistors.
Removed them and re-checked voltages.
Today I finally powered up the whole amplifier to +-58Vdc with the Bulb Tester as my only protection. Blew the 5A fuse when I bypassed the Bulb Tester. It's currently running on a 10A mains fuse until I add in a soft starter. It will run on a T4A with the soft start installed.
The offset VR corrupts the balance of the LTP. Remove it.
Rely on balance of the LTP, resistors and CCS to give good/excellent balance of LTP currents (I have 201mV and 202mV across the two 100ohm LTP degeneration resistors at present - 0.2mV difference across the LTP emitters).
That small error is almost certainly due to the inequality of the re-injected base currents of the CCS and the injection of the VAS/EF base current.
If your not using high quality geniune multi turn pots for ltp adjustment then you will have plorbem. which i had trieed before.
You can always consider putting a 20w bulb-100w to limit maximum power to transistors in case of fault. you can also put a switch inseries to cut off the bulb when confirmed working properly. This nearly saves components from damaged all the time when connection error exist.
20w bulb means 20w max will flow into amp. brighter bulb more power drawed.
In your case, if capacitor bank isn't too big. your amp would likely to be saved when running +-66v and fake transistors.
Last edited:
Hi HI-FI PRO,
For adjustments, you should be using a single turn control. If you must restrict the range of authority, use fixed resistors to restrict the range. Multi-turn controls are not designed to pass current through the wiper! They are for potentiometer use only. So called "cheap" single turn controls are anything but, and most have enough wiper contact to pass modest amounts of current.
As for DC offset, Andrew is quite correct. The match on your diff pair and degeneration resistors (general case here) should be sufficient to keep DC offset voltages low enough to be of no concern. Throwing the diff pair balance off to correct DC offset is not the right way to do things. That's assuming that the design of the amplifier wants to sit midpoint (near audio common) rather than one supply rail or the other.
-Chris
For adjustments, you should be using a single turn control. If you must restrict the range of authority, use fixed resistors to restrict the range. Multi-turn controls are not designed to pass current through the wiper! They are for potentiometer use only. So called "cheap" single turn controls are anything but, and most have enough wiper contact to pass modest amounts of current.
As for DC offset, Andrew is quite correct. The match on your diff pair and degeneration resistors (general case here) should be sufficient to keep DC offset voltages low enough to be of no concern. Throwing the diff pair balance off to correct DC offset is not the right way to do things. That's assuming that the design of the amplifier wants to sit midpoint (near audio common) rather than one supply rail or the other.
-Chris
From my epxrience, tried using single turn pot, fake quality not so good. Even with resistors in parallel and very low load. Pot resistance changes all the time.
Used a multi turn high quality pot in place no more plorbems, no resistor parallel in pot required either.
This is for current adjust of Diffrential Pair.
EDIT: I actually have to question the quality. It works fine in other aplications like adjusting lm317 voltage, but not good as LTP current adjustment.
Used a multi turn high quality pot in place no more plorbems, no resistor parallel in pot required either.
This is for current adjust of Diffrential Pair.
EDIT: I actually have to question the quality. It works fine in other aplications like adjusting lm317 voltage, but not good as LTP current adjustment.
Last edited:
Tapatalk Cloud - Downlaoad File download_20170524_113557.png
Tapatalk Cloud - Downlaoad File download_20170524_113523.png
Tapatalk Cloud - Downlaoad File download_20171003_131013.png
Tapatalk Cloud - Downlaoad File download_20170524_113606.png
Tapatalk Cloud - Downlaoad File download_20170524_113523.png
Im not sure of this will help anyone buy these are some layout photos for my case.
Tapatalk Cloud - Downlaoad File download_20170524_113523.png
Tapatalk Cloud - Downlaoad File download_20171003_131013.png
Tapatalk Cloud - Downlaoad File download_20170524_113606.png
Tapatalk Cloud - Downlaoad File download_20170524_113523.png
Im not sure of this will help anyone buy these are some layout photos for my case.
Hello HI-FI PRO,
I can say with 100% certainty after spending >40 years repairing equipment and >20 years repairing test equipment, for use in audio equipment the single turn control out performs multi-turn controls. The only reason multi-turn controls even exist is to finely adjust things like zero and span in measuring equipment or in process control. But there the controls are used correctly - as potential dividers with almost zero current flowing through the wiper to track. Single turn controls are always used when there is any expectation of current flow in the wiper circuit. Used properly (and with normal industrial quality levels) each type of control will last decades normally.
I've designed a fair amount of stuff over the years too. Most of the stuff I designed and built in high school (secondary school) is still functioning well. The later designs are all still functional. Equipment I see that use these controls outside of their intended purpose typically generate faults related to those problems.
Now, if you want to do things incorrectly for yourself (only), I don't care. However, do not pass on or recommend your practices to anyone else. I do highly recommend that you do something that most folks do not do. Why not read the data sheets on parts you are using? Potentiometer means zero wiper current, rheostat means the wiper can carry current. Read data sheets from manufacturers of good products. You will see that I am correct about this.
It never ceases to amaze me how long different parts can remain working when mis-used. They eventually do fail.
-Chris
Then why do you buy "fake", low quality products in the first place???! Given how much is at stake with a bias control, doesn't it make sense to use a normal quality product? If you don't, your experiences are something you worked hard for and the reason completely escapes me.
That isn't how you use them. Padding resistors go in series with the pot, not in parallel.
No, your problems are just shifted to the future as you are not using the multi-turn control correctly, and that is just for starters. If you are willing to purchase high quality multi-turn controls, why on earth don't you buy high quality single turn controls and compare those?
I can say with 100% certainty after spending >40 years repairing equipment and >20 years repairing test equipment, for use in audio equipment the single turn control out performs multi-turn controls. The only reason multi-turn controls even exist is to finely adjust things like zero and span in measuring equipment or in process control. But there the controls are used correctly - as potential dividers with almost zero current flowing through the wiper to track. Single turn controls are always used when there is any expectation of current flow in the wiper circuit. Used properly (and with normal industrial quality levels) each type of control will last decades normally.
I've designed a fair amount of stuff over the years too. Most of the stuff I designed and built in high school (secondary school) is still functioning well. The later designs are all still functional. Equipment I see that use these controls outside of their intended purpose typically generate faults related to those problems.
Now, if you want to do things incorrectly for yourself (only), I don't care. However, do not pass on or recommend your practices to anyone else. I do highly recommend that you do something that most folks do not do. Why not read the data sheets on parts you are using? Potentiometer means zero wiper current, rheostat means the wiper can carry current. Read data sheets from manufacturers of good products. You will see that I am correct about this.
It never ceases to amaze me how long different parts can remain working when mis-used. They eventually do fail.
-Chris
Even with resistors in parallel and very low load.
That isn't how you use them. Padding resistors go in series with the pot, not in parallel.
When resistors in parallel, this means when pot in operation, should it be low quality and suddenly open circuit for milli seconds, there will still be connection in circuit, not a open circuit. This can possibly save a amp from blowing up.
Then why do you buy "fake", low quality products in the first place???!
I thought they were real parts and would work, so I used them.
However managed to get multi turn pots like 13turns on aliexpress for 0.8-1usd shipping included and 10pices pots. They have a gold adjustment metal piece and worked very well. I would recomend them because I used them in more than 2 amplifiers and worked very well.
If you are willing to purchase high quality multi-turn controls, why on earth don't you buy high quality single turn controls and compare those?
I'd say Multi turn vs single is just personall preference and application dependant.
The stuff I adjust in my amp is always super precison, multi turn always prefered.
But there the controls are used correctly - as potential dividers with almost zero current flowing through the wiper to track. Single turn controls are always used when there is any expectation of current flow in the wiper circuit. Potentiometer means zero wiper current
You are not correct,
http://docs-asia.electrocomponents.com/webdocs/087d/0900766b8087d7ed.pdf
http://docs-asia.electrocomponents.com/webdocs/0fb6/0900766b80fb659f.pdf
Heres two diffrent pots i quickly searched up on rs components.
The single turn is rated for 50mA only, while the multi turn is rated for 100mA.
Pots are power dependant as long as you do not exceed max current.
No current flowing through pots do not exist when used in any circuit operation.
For some sort of operation current and voltage will always exist/flow through the pot, its only a matter of ammount of voltage and current.
That isn't how you use them. Padding resistors go in series with the pot, not in parallel.
When resistors in parallel, this means when pot in operation, should it be low quality and suddenly open circuit for milli seconds, there will still be connection in circuit, not a open circuit. This can possibly save a amp from blowing up.
Then why do you buy "fake", low quality products in the first place???!
I thought they were real parts and would work, so I used them.
However managed to get multi turn pots like 13turns on aliexpress for 0.8-1usd shipping included and 10pices pots. They have a gold adjustment metal piece and worked very well. I would recomend them because I used them in more than 2 amplifiers and worked very well.
If you are willing to purchase high quality multi-turn controls, why on earth don't you buy high quality single turn controls and compare those?
I'd say Multi turn vs single is just personall preference and application dependant.
The stuff I adjust in my amp is always super precison, multi turn always prefered.
But there the controls are used correctly - as potential dividers with almost zero current flowing through the wiper to track. Single turn controls are always used when there is any expectation of current flow in the wiper circuit. Potentiometer means zero wiper current
You are not correct,
http://docs-asia.electrocomponents.com/webdocs/087d/0900766b8087d7ed.pdf
http://docs-asia.electrocomponents.com/webdocs/0fb6/0900766b80fb659f.pdf
Heres two diffrent pots i quickly searched up on rs components.
The single turn is rated for 50mA only, while the multi turn is rated for 100mA.
Pots are power dependant as long as you do not exceed max current.
No current flowing through pots do not exist when used in any circuit operation.
For some sort of operation current and voltage will always exist/flow through the pot, its only a matter of ammount of voltage and current.
Okay HI-FI PRO,
Looks like I'll have to look up the information for you from the manufacturer. I really wish you guys would research parts for yourself instead of having someone else do it for you.
Vishay : http://www.vishay.com/docs/51001/potentiom.pdf Read section 1.1 A voltage divider by definition does not have appreciable current flow. Lower than 1 mA. The rest is a quick read. This information applies to the vast majority of multi-turn controls you will find out in the wild. The use of cermet in a multi-turn control is a new thing. I do thank you for bringing this up. However, this type of control is definitely not the standard part!
Bourns:
http://www.bourns.com/data/global/pdfs/ap_proc.pdf Here look under ""The Resistive Element". They seem to be singling out the wire wound, multi-turn control for wiper current and not so much the Cermet option. No mention is made of the carbon film types that were popular in earlier times. Those most certainly did not like wiper current flow!
As I said before, generally speaking, multi-turn controls are for potentiometer applications. They do not have the contact area to support current flow long term. They may work, but they sure aren't designed to. If you have any doubts, use your own eyeballs and take a multi-turn control apart and look at it. Then do the same with a single turn control. Common sense is all you need to have here.
There is nothing in an amplifier that will stay put as far as adjustment goes.
Now for the real kicker. I don't know who "multi turn always prefered" is, but a real designer restricts the range of the single turn control by using resistors in series. A single turn control can give you an amazingly fine range of adjustment. Lazy designers use multi-turn controls because they haven't figure this out yet.
The power dissipated in a shorter section of resistive element must be derated from the maximum.
Although you did find one data sheet that supports your position that multi-turn controls can stand up to wiper current, it is an exception and not the norm. Best practices would include resistance padding the control at both ends to define the maximum and minimum range of adjustment. This practice automatically allows a single turn control to have a reasonable range of adjustment while also allowing close adjustments for whatever parameter you are adjusting for. Please give this some consideration as you cannot trust that people will purchase the specific multi-turn control that will take current flow through the wiper. Most controls do not. Not at this point in time anyway.
-Chris
Looks like I'll have to look up the information for you from the manufacturer. I really wish you guys would research parts for yourself instead of having someone else do it for you.
Vishay : http://www.vishay.com/docs/51001/potentiom.pdf Read section 1.1 A voltage divider by definition does not have appreciable current flow. Lower than 1 mA. The rest is a quick read. This information applies to the vast majority of multi-turn controls you will find out in the wild. The use of cermet in a multi-turn control is a new thing. I do thank you for bringing this up. However, this type of control is definitely not the standard part!
Bourns:
http://www.bourns.com/data/global/pdfs/ap_proc.pdf Here look under ""The Resistive Element". They seem to be singling out the wire wound, multi-turn control for wiper current and not so much the Cermet option. No mention is made of the carbon film types that were popular in earlier times. Those most certainly did not like wiper current flow!
As I said before, generally speaking, multi-turn controls are for potentiometer applications. They do not have the contact area to support current flow long term. They may work, but they sure aren't designed to. If you have any doubts, use your own eyeballs and take a multi-turn control apart and look at it. Then do the same with a single turn control. Common sense is all you need to have here.
... and you would be wrong.
Then you have designed the circuit incorrectly. Normally when a bias control goes open, current falls to a minimum, or zero depending on the circuit. It is possible to design a bias circuit that failure of the control causes maximum current flow. Several designers have made this basic error.
That was a mistake on your part. You compounded that mistake by trying to compare the cheap, fake single controls to properly manufactured multi-turn controls. This is the error I was pointing out.
Show me a precision adjustment in an amplifier! Bias current drifts around depending on temperature, run time, breezes, AC voltage coming in and probably dust load on the heat sink. DC offset is similarly imprecise as it can also wander around on its own. If you are adjusting SOA limiting, that is also imprecise.
There is nothing in an amplifier that will stay put as far as adjustment goes.
Now for the real kicker. I don't know who "multi turn always prefered" is, but a real designer restricts the range of the single turn control by using resistors in series. A single turn control can give you an amazingly fine range of adjustment. Lazy designers use multi-turn controls because they haven't figure this out yet.
Potentiometers conduct uA or nA in operation. Never even 1 mA.
Those ratings are atypical for a multi-turn control. Good find, but normal controls are not capable of passing current through the wiper.
Prepare yourself for a nasty surprise some day.
The power dissipated in a shorter section of resistive element must be derated from the maximum.Although you did find one data sheet that supports your position that multi-turn controls can stand up to wiper current, it is an exception and not the norm. Best practices would include resistance padding the control at both ends to define the maximum and minimum range of adjustment. This practice automatically allows a single turn control to have a reasonable range of adjustment while also allowing close adjustments for whatever parameter you are adjusting for. Please give this some consideration as you cannot trust that people will purchase the specific multi-turn control that will take current flow through the wiper. Most controls do not. Not at this point in time anyway.
-Chris
Show me a precision adjustment in an amplifier!
Sorry but I cannot go into detail about that, becasue that would mean going into my amplifier design principles and introduce competetors if I decide to sell amps.
Then you have designed the circuit incorrectly. Normally when a bias control goes open, current falls to a minimum, or zero depending on the circuit. It is possible to design a bias circuit that failure of the control causes maximum current flow. Several designers have made this basic error.
I don't use vbe multipliers in my amp designs to bias, sometimes in other applications theres no way you can make it so it can open circuit and still be fine.
Potentiometers conduct uA or nA in operation. Never even 1 mA.
With refrence to amp schematic, you see it says 3.75mA accross r14. Collector current of q7 is also approx equal to emitter current of q7. And current law, sum of currents entering equal to leaving. That means 3.75mA is passed through r7 ccs adjust 200r.
Sorry but I cannot go into detail about that, becasue that would mean going into my amplifier design principles and introduce competetors if I decide to sell amps.
Then you have designed the circuit incorrectly. Normally when a bias control goes open, current falls to a minimum, or zero depending on the circuit. It is possible to design a bias circuit that failure of the control causes maximum current flow. Several designers have made this basic error.
I don't use vbe multipliers in my amp designs to bias, sometimes in other applications theres no way you can make it so it can open circuit and still be fine.
Potentiometers conduct uA or nA in operation. Never even 1 mA.
With refrence to amp schematic, you see it says 3.75mA accross r14. Collector current of q7 is also approx equal to emitter current of q7. And current law, sum of currents entering equal to leaving. That means 3.75mA is passed through r7 ccs adjust 200r.
I did a quick search on rs components, virtually all multi turn pots on rs use cermet as element material.
MUFF WIGGLER :: View topic - Ident-a-pot Vol. 2 (Bourns Trimmer - fake or not)
Since I talked about pots on aliexpress (1usd for 10 pices), i'd share this page on others takling about chiense pots.
The ones I own and had used without any plorbem are branded baoter and bochen. The bochens I own were more epxensive and are 26 turns, while baoters are only 13 turns.
Last edited:
With refrence to amp schematic, you see it says 3.75mA accross r14. Collector current of q7 is also approx equal to emitter current of q7. And current law, sum of currents entering equal to leaving. That means 3.75mA is passed through r7 ccs adjust 200r.
Dear Mr PRO, I am an ardent follower of your divine wisdom of high-end audio amplifier circuits who is trying to learn from you, perhaps I am a bit challenged and didn't understand fully due to mild dementia, but does your statement above mean that the current through R17 DC-offset trimmer as seen in the attached schematics is also 3.75 mA?
Regards
HI-FI PRO,
Your reliance on multi-turn controls shows a lack of engineering skill as well. Maybe after a decade or two you might learn how to use single turn controls successfully. If you would observe commercial products, very few utilise multi-turn controls. The reason is not that they are less expensive.
-Chris
Oh please, get real! The only people who we've heard that from before are real head cases. But there is the thing about selling a product. As soon as you sell the first one, your secrets are out as they can be reverse engineered, and that is more valuable than a correct schematic. The very thought that you can design something that hasn't already been done before is pretty sad. Any patents you file will either fail on prior art, or be granted because there is something very wrong with the design. The saddest outcome would be a failed patent attempt due to prior art that was never produced because it doesn't work well.
You're going to need a non-linear element is you hope to have bias that tracks with temperature. Maybe a diode clamp of some type for protection, given that mere mortals are not able to understand your design and therefore cannot give you an exact answer.
Your reliance on multi-turn controls shows a lack of engineering skill as well. Maybe after a decade or two you might learn how to use single turn controls successfully. If you would observe commercial products, very few utilise multi-turn controls. The reason is not that they are less expensive.
-Chris
Thanks for following
The current through r17 dc off set trimmer is not 3.75mA. To annalyse this we would assume pot is centered and 500ohms both side.
Effectivly there are two 100ohms parallel with the 500ohms going to each side of transistor.
we also assume transitor of diffrential pair share load equaly, meaning q1 has 1.875mA accross it (3.75/2)
This also means 1.875mA is accross r15(100ohm)parallel with r17(500ohm).
You can calculate current divider or method you know.
(100/600)(1.875) = 0.31125mA accross r17 when it is centered (500ohm both side pot).
Your reliance on multi-turn controls shows a lack of engineering skill as well
No, because my amplifier schematic can actually go that precise with the driver biasing to get you extra sound quality
The very thought that Chinese are never supposed to be innovative is just sad and racist. I have many refrence amps including mid price tube amps too, the transistor amps I designed smoke them.
Last edited:
Hi HI-FI PRO,
If you are such a great designer with superior products such as you say, why haven't we heard of you before? Also, if this is true, why are your wrong ideas so blatant? Apparently you have ego to boot, but are suffering from a lack of engineering chops. I just read your response in another thread to a class A design where you demonstrate a lack of comprehension for the basics of design. But hey, that's your world. We don't need to experience it so often.
Next time you try and play the race card, pick someone who might be racist. My friends and colleagues come from every corner of the world. You're barking up the wrong tree.
-Chris
Now that is funny! You do understand that bias varies as the load impedance changes, and speakers are not even close to having a flat impedance. You're funny, I think I like you.
Given that I don't know where you actually come from, I think you are reaching here.
If you are such a great designer with superior products such as you say, why haven't we heard of you before? Also, if this is true, why are your wrong ideas so blatant? Apparently you have ego to boot, but are suffering from a lack of engineering chops. I just read your response in another thread to a class A design where you demonstrate a lack of comprehension for the basics of design. But hey, that's your world. We don't need to experience it so often.
Next time you try and play the race card, pick someone who might be racist. My friends and colleagues come from every corner of the world. You're barking up the wrong tree.
-Chris
Replacing the driver pair MJE15032/MJE15033 with original parts didn't make much difference. Q20 overheated and blew F2. I noticed that unlike other power stage transistors Q20 started heating up quite early, even at idle. It blew shortly after I added 6Ohm load. Other power transistors were cool, with maybe exception of Q21 which was luke-worm.
Not sure what the next step should be. Replace just Q20 with a new part and see if it acts the same, or if a difference transistor overheats now? Replace all power transistors (I currently use NJW0281 / NJW0302) with something different to see if that makes a difference?
Not sure what the next step should be. Replace just Q20 with a new part and see if it acts the same, or if a difference transistor overheats now? Replace all power transistors (I currently use NJW0281 / NJW0302) with something different to see if that makes a difference?
When my designs fully warm up it stays at a fix bias. We then set to perfection, it will say at this bias point and even with +-20C in room temps will still stay near optimum operating point for gain stages. We are talking about voltage gain stages, not output stage bias current.Hi HI-FI PRO,
Now that is funny! You do understand that bias varies as the load impedance changes, and speakers are not even close to having a flat impedance. You're funny, I think I like you.
If you are such a great designer with superior products such as you say, why haven't we heard of you before?
-Chris
I don't have any amplifier porducts selling yet. I plan to sell in the future.
Given that I don't know where you actually come from, I think you are reaching here.
I thought it was obvious, since my diy audio profile has a Chinese flag.
Last edited:
Possibly q20 was leaky and damaged from previous tests. Power up the amp with just one pair of outputs and test(you can use any cheap transistors to test, just make sure voltage rating is at least 120v). Put some music on mid low level of volumes too, if it works then should be good, go to 3 pairs of outputs