cjd said:
Can you send a link to the schematic you are referencing?
My schematic is nothing like that.
I have 2K resistor from input to base of each transistor, and 22K to GND. There is a 100K resistor after the 2K resistor to the LM3886 output.
Cheers!
Russ
Check this post for the schematic at the core of my amp. R3 and R9 are the two I speak of - the resistors right at the input of the chip, not at the input of the circuit itself.
Since this is a bootstrapped setup, I adapted/mirrored this to put a single LM4780 on each side, paralleled (R0.1 on each output per data sheet). I have LM3886 boards here I can work with if this seems likely to be the issue, but I mis-ordered and don't have big caps for the chip - could toss some 220's on though. Should work with the regulated power supply (10k uF on that per rail).
I know there are a couple additions beyond the schematic I linked to which I have captured in my amp.
I have the same problems with this amp being both very quiet and going into distortion mode beyond a certain point when I'm running the X-BOSOZ into the JT into the X-GC. Though I had that 100K pot in front of the X-BOSOZ as a safety point. I can take it out of the circuit to be absolutely sure.
C
Since this is a bootstrapped setup, I adapted/mirrored this to put a single LM4780 on each side, paralleled (R0.1 on each output per data sheet). I have LM3886 boards here I can work with if this seems likely to be the issue, but I mis-ordered and don't have big caps for the chip - could toss some 220's on though. Should work with the regulated power supply (10k uF on that per rail).
I know there are a couple additions beyond the schematic I linked to which I have captured in my amp.
I have the same problems with this amp being both very quiet and going into distortion mode beyond a certain point when I'm running the X-BOSOZ into the JT into the X-GC. Though I had that 100K pot in front of the X-BOSOZ as a safety point. I can take it out of the circuit to be absolutely sure.
C
Ok, I was not sure how you had the input into the amp arranged.
I can tell you that the schematic is correct. My amp is nearly identical. My resistor values for what is R4 and R10(and their opposites) on that schematic are 10K and 499ohms, and mine is working perfectly, so I would not suspect what you have should be too far off.
I can tell you that the schematic is correct. My amp is nearly identical. My resistor values for what is R4 and R10(and their opposites) on that schematic are 10K and 499ohms, and mine is working perfectly, so I would not suspect what you have should be too far off.
I figured it was correct but... I guess I'll give it a shot without the 100k first, and then with the 3886 boards if I still have the issue with the 100k missing.
I should probably mention, I had anticipated experimenting with some of CarlosFM's RC additions and whatnot, and just assembled with them in place. I don't know that those would cause a problem but... Let me see if I can get a readable version of the exact schematic I'm using posted up here.
C
I should probably mention, I had anticipated experimenting with some of CarlosFM's RC additions and whatnot, and just assembled with them in place. I don't know that those would cause a problem but... Let me see if I can get a readable version of the exact schematic I'm using posted up here.
C
cjd,
I did the back of the envelope calculation, and the 100k pot is very, very likely the cause of your woes. Because of the low input impedance of the amp circuit, a -2dB attenuation setting on the 100k pot is actually a -21dB attentuation when you include the input impedance.
Cheers, Terry
I did the back of the envelope calculation, and the 100k pot is very, very likely the cause of your woes. Because of the low input impedance of the amp circuit, a -2dB attenuation setting on the 100k pot is actually a -21dB attentuation when you include the input impedance.
Cheers, Terry
It's actually really hard to say. I'm not sure that the source I'm playing with that cranks out the juice to get this louder isn't itself the problem (actually, I suspect it may be a lot of it), and I'm just causing myself trouble using that to go loud instead of tracking down why the circuit itself isn't doing a whole lot of amplification.
CD player to X-BOSOZ (TP kit) to Joshua Tree to X-GC and I definitely never get into distortion, but I also never get much output. I can crank it all the way up and sit there a couple feet away from the speaker and not think it's even slightly loud (I'm guessing mid 80dB levels).
I'll get some more specific measurements later if I have time (voltage as well as dB).
I also have one JT channel that has a short somewhere - it starts out at about 2/3 volume. Past a certain point it does get louder. The other one has a nifty pop at one switch but works the full range.
Over the next few days I'll get another board out and wire up an X-GC with a pair of 3886 chips instead of the pair of 4780's.
C
CD player to X-BOSOZ (TP kit) to Joshua Tree to X-GC and I definitely never get into distortion, but I also never get much output. I can crank it all the way up and sit there a couple feet away from the speaker and not think it's even slightly loud (I'm guessing mid 80dB levels).
I'll get some more specific measurements later if I have time (voltage as well as dB).
I also have one JT channel that has a short somewhere - it starts out at about 2/3 volume. Past a certain point it does get louder. The other one has a nifty pop at one switch but works the full range.
Over the next few days I'll get another board out and wire up an X-GC with a pair of 3886 chips instead of the pair of 4780's.
C
So i turn on the first channel. Regulated PSU is Ok, i obtain +/- 26,40V on rail. But Offset is not good. I obtain with + and - In shorted to graound 57mV for -OUT and a awfull 2,2V for +OUT.
Here's the schém and layout If you have idees where to search?
Don't pay attention to value on layout they the old, goods one are on schématic.
Marc
Here's the schém and layout If you have idees where to search?
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Don't pay attention to value on layout they the old, goods one are on schématic.
Marc
I don't think that I can help you with your trouble, but I do notice a couple things.
Do you have the resistors from inputs to ground that were to the schematic fairly recently? There was one other thing added, 500R somewhere. In the schematic I posted, R11/R12 and R6.
Also, it is generally recommended that the feedback resistor be very close to the chip (if not attached to the pins directly) - that would be R13/R14 in your schematic.
C
Do you have the resistors from inputs to ground that were to the schematic fairly recently? There was one other thing added, 500R somewhere. In the schematic I posted, R11/R12 and R6.
Also, it is generally recommended that the feedback resistor be very close to the chip (if not attached to the pins directly) - that would be R13/R14 in your schematic.
C
cjd said:
My work is based on metalman schématic and i do not see any res between in and ground it's the first time i see that. What the function of these!?
I do somme mesurment in desoldering feedback resistor (by me 49K9 R7/R8) and what I could said is when the chip stage is connected the paire diff stage shown some diff. Before the interstage cap the voltage is 15v for the side with 2V2 offset. The voltage across colector res from the diff paire is different too : 6V for the 2V2 offset side and 12v for the other. When i desolder coupling cap, the 2 collector res voltage is the same with about 9V. Resoldering the coupling cap and the voltage are the same as before.
For all ZTX450 i can measure a 0V6 VBE.
Marc
Marc,
The addition of 20k resistors from input to ground is specifically intended to eliminate DC offset. When I originally designed the circuit, the bias current trhough the input bipolar transistors was relatively small ( a few milliamps), so the transistor base pins were requiring only a microamps of current which they could easily draw through the existing resistor networks (such as R7+R13+R11 on your schematic). After discovering that the input transistors were prematurely clipping, I increased the bias on each transistor to about 20mA, which creates a current draw at the base of roughly 0.1mA, which the circuit tries to draw from ground through the high value resistor network, but finds an easier current source from teh op-amp outputs causing the DC offset. Thus I added the 20k resistors between the input and ground to provide a better ground reference. This is also why the you get low offset on the side where you have shorted the input to ground.
If the ouputs of your preamp are ground referenced, then the offset will only occur when the amp is not connected to your pre, and will dissappear when you connect to the preamp. Otherwise adding the extra resistors will solve the problem.
All of this was a fairly late occuring discovery, which I documented HERE . I hope it doesn't inconvenience you too much.
Terry
The addition of 20k resistors from input to ground is specifically intended to eliminate DC offset. When I originally designed the circuit, the bias current trhough the input bipolar transistors was relatively small ( a few milliamps), so the transistor base pins were requiring only a microamps of current which they could easily draw through the existing resistor networks (such as R7+R13+R11 on your schematic). After discovering that the input transistors were prematurely clipping, I increased the bias on each transistor to about 20mA, which creates a current draw at the base of roughly 0.1mA, which the circuit tries to draw from ground through the high value resistor network, but finds an easier current source from teh op-amp outputs causing the DC offset. Thus I added the 20k resistors between the input and ground to provide a better ground reference. This is also why the you get low offset on the side where you have shorted the input to ground.
If the ouputs of your preamp are ground referenced, then the offset will only occur when the amp is not connected to your pre, and will dissappear when you connect to the preamp. Otherwise adding the extra resistors will solve the problem.
All of this was a fairly late occuring discovery, which I documented HERE . I hope it doesn't inconvenience you too much.
Terry
metalman said:
Hi terry, Ok for the 20k input to ground. I will test. Just a remark. When I mesure the offset all inpiut are shorted to ground (+IN and -IN)
Marc
Ok, more information.
Once again:
Source > X-BOSOZ > Joshua Tree > X-GC
I fired it up again tonight to measure some stuff - I get under 1V at the outputs (~85dB at 1M) which is about expected - get about half that at O+ or O- to G. Each individual chip-amp board works properly out of this circuit and in (wiring speaker O+ or O- to G) So I'm confident at this point the circuit is itself all there. This was with a LM4780 on each side, the two halves paralleled.
Next, I swapped out the 4780 boards for 3886 boards. This will (theoretically) result in the exact circuit Terry has posted. Not surprisingly, I picked up a few dB (each chip is seeing half the impedance - in this case, 4ohm, which is where this is probably ideal - my mains dip below 3ohm, which would put 1.4ohm per chip on this circuit so I need the 4780's), but overall the same thing. Very quiet with the volume turned all the way up. And in this case, on some bass-heavier tracks I'm definitely hearing signs of distortion.
My next thought is to bypass the power supply boards (regulated) - the only trouble here is that I don't have a variac and don't have an appropriately sized transformer. However, these are putting out a rock steady 31.6V (which is exactly what the calculation says it should be) so I very much doubt this is the issue. But it could be.
I'm lost.
What should I be looking at next? I will try to remember to post up the exact PCB's I used for this, with the hand-wired parts (i.e. not on the PCB) illustrated or noted.
C
Once again:
Source > X-BOSOZ > Joshua Tree > X-GC
I fired it up again tonight to measure some stuff - I get under 1V at the outputs (~85dB at 1M) which is about expected - get about half that at O+ or O- to G. Each individual chip-amp board works properly out of this circuit and in (wiring speaker O+ or O- to G) So I'm confident at this point the circuit is itself all there. This was with a LM4780 on each side, the two halves paralleled.
Next, I swapped out the 4780 boards for 3886 boards. This will (theoretically) result in the exact circuit Terry has posted. Not surprisingly, I picked up a few dB (each chip is seeing half the impedance - in this case, 4ohm, which is where this is probably ideal - my mains dip below 3ohm, which would put 1.4ohm per chip on this circuit so I need the 4780's), but overall the same thing. Very quiet with the volume turned all the way up. And in this case, on some bass-heavier tracks I'm definitely hearing signs of distortion.
My next thought is to bypass the power supply boards (regulated) - the only trouble here is that I don't have a variac and don't have an appropriately sized transformer. However, these are putting out a rock steady 31.6V (which is exactly what the calculation says it should be) so I very much doubt this is the issue. But it could be.
I'm lost.
What should I be looking at next? I will try to remember to post up the exact PCB's I used for this, with the hand-wired parts (i.e. not on the PCB) illustrated or noted.
C
In post 528, I put the wrong schema. The good one is this under.
Marc
An externally hosted image should be here but it was not working when we last tested it.
Marc
I do some more investigation on the 2,2V offset LM. Coupling Cap and feedback resistor are desoldered. I can mesure a 0,350 ground to LMinput pin voltage amount and a 2,3 ground to LMoutputpin voltage amount. It's looks like the Lm present trouble. I will relace the wall card by the second to see what i can mesure. If results are normal the LM is in fault.
Marc
Marc
CJD
Are your input leads reversed?
I had a similar, low output experience with input leads connected wrong on a different amp.
Are your input leads reversed?
I had a similar, low output experience with input leads connected wrong on a different amp.
Re: CJD
How does it make a difference? I'll gladly give this a shot as it's exceptionally easy to swap. But as far as I can tell, there is nothing here to say one side is this way or that - the circuit is electrically a mirror image... or so I see it (which could be the trouble).
I'll report back later on this. 🙂 I get to fire it up for a few minutes every night for no reason other than I have to water a whole mess of plants in my workshop as they wait the weather to turn so we can plant 'em outside. 🙂 mmmm, fresh garden produce!
C
Ed Lafontaine said:
How does it make a difference? I'll gladly give this a shot as it's exceptionally easy to swap. But as far as I can tell, there is nothing here to say one side is this way or that - the circuit is electrically a mirror image... or so I see it (which could be the trouble).
I'll report back later on this. 🙂 I get to fire it up for a few minutes every night for no reason other than I have to water a whole mess of plants in my workshop as they wait the weather to turn so we can plant 'em outside. 🙂 mmmm, fresh garden produce!
C
Hey, anyone contributing is OK in my book. 🙂
I did test it and it didn't change anything. Ahh well. Good thought anyhow Ed.
Drilled about 16 holes and got two tapped. I think I have another 16 holes to drill and tap beyond that, plus 5 more for the X-BOSOZ - maybe another 4 beyond that. Slooow work. So, I really really want to get this working! 🙂
The reason I keep going back to those 10k resistors in series with the chip input is that, well... that's ten times what the datasheet puts there - and each individual chip (single ended, no X circuit) in this configuration is itself very quiet.
Is there gain expected in the X independant of the amp circuit?
What happens if I lower this value? (Rb1-4 in my schematic)
C
I did test it and it didn't change anything. Ahh well. Good thought anyhow Ed.
Drilled about 16 holes and got two tapped. I think I have another 16 holes to drill and tap beyond that, plus 5 more for the X-BOSOZ - maybe another 4 beyond that. Slooow work. So, I really really want to get this working! 🙂
The reason I keep going back to those 10k resistors in series with the chip input is that, well... that's ten times what the datasheet puts there - and each individual chip (single ended, no X circuit) in this configuration is itself very quiet.
Is there gain expected in the X independant of the amp circuit?
What happens if I lower this value? (Rb1-4 in my schematic)
C