Gilmore Dynalo troubleshooting (Desperate)
I've decided to take my trouble here now instead of just AMB's help website, but this amp is giving me to much grief. I am absolutely out of (my known) options.
You can get the schematic here for the headphone amp: http://www.djgardner.com/headphone/gilmore/dynalo/layout/Rev-C/dynalo_rev_c.pdf
Here is the help thread on AMB's forum (ignore the first couple of posts, the rest are significant): AMB Laboratories DIY Audio • View topic - High DC offset in the right channel
Now for some details:
The left channel works perfectly fine, the offset always stays below 5.0mV and it doesn't do any crazy offset spikes like the right channel does.
I've installed a zobel network on both channels consisting of a 22ohm resistor and a 0.047uf under AMB's advisory, but that hasn't fixed the constant switching of offset of ~5.0mV to ~400.0mV every couple of minutes.
I currently have the gain of the amp set to 3x (R16=2k R7=1K C27=33pF)
Any form of help would be massive at this point, This amp is driving (har har) me nuts.
Thanks,
Kurt
I've decided to take my trouble here now instead of just AMB's help website, but this amp is giving me to much grief. I am absolutely out of (my known) options.
You can get the schematic here for the headphone amp: http://www.djgardner.com/headphone/gilmore/dynalo/layout/Rev-C/dynalo_rev_c.pdf
Here is the help thread on AMB's forum (ignore the first couple of posts, the rest are significant): AMB Laboratories DIY Audio • View topic - High DC offset in the right channel
Now for some details:
The left channel works perfectly fine, the offset always stays below 5.0mV and it doesn't do any crazy offset spikes like the right channel does.
I've installed a zobel network on both channels consisting of a 22ohm resistor and a 0.047uf under AMB's advisory, but that hasn't fixed the constant switching of offset of ~5.0mV to ~400.0mV every couple of minutes.
I currently have the gain of the amp set to 3x (R16=2k R7=1K C27=33pF)
Any form of help would be massive at this point, This amp is driving (har har) me nuts.
Thanks,
Kurt
Last edited:
Hi,
I've seen the schematic and read the communication between AMB and
you.AMB is expecting you know what you are doing but apparently you are not.From the high DC offset of the right channel tells you something is very
wrong somewhere.
1.Well for starters basic troubleshooting in DIY is to check
against the good channel whether all resistors and components are mounted
correctly and are of the said values in schematic.
2.Measure all relevant voltage drops to see if there are any huge discrepencies between the good channel.This will help to zero in on the
possible section of the circuit.
3.Check all pots to see if they are working and functional.
4.Lastly if you suspect oscillation then a scope will be useful.
Good luck. singa
I've seen the schematic and read the communication between AMB and
you.AMB is expecting you know what you are doing but apparently you are not.From the high DC offset of the right channel tells you something is very
wrong somewhere.
1.Well for starters basic troubleshooting in DIY is to check
against the good channel whether all resistors and components are mounted
correctly and are of the said values in schematic.
2.Measure all relevant voltage drops to see if there are any huge discrepencies between the good channel.This will help to zero in on the
possible section of the circuit.
3.Check all pots to see if they are working and functional.
4.Lastly if you suspect oscillation then a scope will be useful.
Good luck. singa
Thank you singa and artu
1. I've checked that a multitude of times, I really wish there was a schematic showing what all the voltage drops should be on the board, I'll spend some time today and check all the voltage drops on the board, I'll post them here for reference.
2. Same as above, except the left channel will become the reference channel
3. Only one pot at the moment, and that's the log volume pot. That pot is working fine. I'll install any of the DC adjust when this whole ordeal is done.
4. I'll go by my electronics lab today sometime after lunch, I'll borrow a scope and see if i can (find oscillation) sort out the problem.
The DC servo is installed in the left channel at the moment, when the amp gets stable, I'll be doing all the adjusting necessary.
Thanks!
1. I've checked that a multitude of times, I really wish there was a schematic showing what all the voltage drops should be on the board, I'll spend some time today and check all the voltage drops on the board, I'll post them here for reference.
2. Same as above, except the left channel will become the reference channel
3. Only one pot at the moment, and that's the log volume pot. That pot is working fine. I'll install any of the DC adjust when this whole ordeal is done.
4. I'll go by my electronics lab today sometime after lunch, I'll borrow a scope and see if i can (find oscillation) sort out the problem.
The DC servo is installed in the left channel at the moment, when the amp gets stable, I'll be doing all the adjusting necessary.
Thanks!
I took a video of the scope when I was moving the volume pot, the steady line is the left channel, the line that's constantly moving up an down is the right channels output.
YouTube - Dynalo troubleshooting (scope)
I really scared that there's a problem with one of the jfets that came from AMB, since I'd have to replace all four.
I haven't measured all the voltage drops, but i did notice that when the offset goes haywire (~-400mV), the voltage drops changed across R18/R20. The voltages on the good board had ~1.45V across both, when the bad board has high offset the values change to about ~1.9V and ~1.1V, could Q3 or Q4 be fried?
Just a note, the offset on the right channel was never this large, It was like a switch had been turned on, and now it has this problem.
Thank you for taking the time to help out, I really really appreciate it at this point.
YouTube - Dynalo troubleshooting (scope)
I really scared that there's a problem with one of the jfets that came from AMB, since I'd have to replace all four.
I haven't measured all the voltage drops, but i did notice that when the offset goes haywire (~-400mV), the voltage drops changed across R18/R20. The voltages on the good board had ~1.45V across both, when the bad board has high offset the values change to about ~1.9V and ~1.1V, could Q3 or Q4 be fried?
Just a note, the offset on the right channel was never this large, It was like a switch had been turned on, and now it has this problem.
Thank you for taking the time to help out, I really really appreciate it at this point.
I took the time to reheat a lot of the solder points on the board, the output transistors in particular, reconnected the power, turned it on... and it became stable although the offset was sitting at around ~20-25mV. I'll need to trim that down. I'm glad I got that worked out, it worries me that the high offset may come again...
Now, the thing that worries me, is when i turn the pot, the offset still moves a good +/-10mV with and without the servo. I'd like a theory as to why, does it have to do with the initial untrimmed offset?
It's nice to know that things are getting better, but I'm still worried, and my lack of electronics knowledge provides me without answers.
Now, the thing that worries me, is when i turn the pot, the offset still moves a good +/-10mV with and without the servo. I'd like a theory as to why, does it have to do with the initial untrimmed offset?
It's nice to know that things are getting better, but I'm still worried, and my lack of electronics knowledge provides me without answers.
Hi, Measure voltage drop across R11,15 (499 ohm) of good left channel
and see if the respective resistors at the right channel are the same ,if
not adjust R10,14 respectively.This should be your starting point reference
before you go on to the rest of the circuit.Also measure voltage across D1,2
led that it is 1.6V as stated.( If the voltage across R11,15 does not change
then measure with respect to common/ground or at the emitter of transistor,
I'm getting rusty at this).
Another thing do measure the good channel voltage at the Jfets and compare
with the faulty right channel.That should give you a good idea about the
situation of the front input section. Good luck. singa
Last edited:
maybe the potentiometer is bad? can you swap channels to see if the problem follows the pot? bad ground to pot? with amp off, measure the resistors . .. on or off, tap components to find an intermittent ? troubleshoot with the feedback loop open (original version has a gain of 35 open loop), it may be easier to find the problem (pull up one side of R16) . . . try circuit cooler (or a hot air gun) to isolate bad component . . . if you think jfets are bad, you could swap them left/right. it seems like several of the components (or connections) could cause this to happen, so it's difficult to isolate. good luck !
.
.
Last edited:
the original version uses component matching to fix offset, wo/adjust potentiometer
.
An externally hosted image should be here but it was not working when we last tested it.
.
Last edited:
double check the vol. pot. with it off, measure the resistance from the input JFET to ground . .. make sure the resistance looks reasonable when you turn the pot. a connection can be bad (esp. to ground), a wiper can be dirty (try contact cleaner?), or the element defective (replace the potentiometer)
.
.
Some headway!
The pot checks out just fine. The there doesn't appear to be any problems with with the resistance of the pot
It almost looks like the DC offset of the other board moves with the pot too, which means it might be the change of impedance... It's just a lot more noticeable on the right board because the initial offset was much higher then the left. Meaning my only way of fixing this is finding a part that should be quite common, but for some reason is not so. AMB outlined this guide here to eliminate the DC offset using R10/R14:
AMB Laboratories DIY Audio • View topic - large DC offset in left channel
When he says turn it to minimum, the pot should be at it's maximum resistance. AMB in his guide states to use a 10K pot, but the schematic shows a 100K pot, I almost think it should be safer to use a 100K pot. I don't quite understand what he means by "Turn one of the two trimpots to change the offset toward 0mV, a little at a time, and simultaneously turn the other trimpot in the opposite direction.". If the pots already at it's maximum resistance, then what's the point?
I'll also attempt to find some moderate sized TO-220 heatsink that i can use for the output transistors. Something like this:
Digi-Key - 294-1042-ND (Manufacturer - PB1-36CB)
or this:
Digi-Key - 294-1080-ND (Manufacturer - 7-340-2PP-BA)
I think i can partially blame it on thermal run away, since the offset will start at something like ~-1.0mV and as the amp warms up, it'll end up at something like ~-25.0mV and the transistors will be about ~85c.
I'll do some more voltage measurements later dealing with the Jfets, I need a break for now.
Thanks a lot!
The pot checks out just fine. The there doesn't appear to be any problems with with the resistance of the pot
The LED's on both boards match out to be 1.65V on all four, same with R11 and R15 I measure exactly 1V drop across them. Apparently my local electronics shop doesn't have any 100k multiturn pots, they could get them in, but they were about $1.25 a turn, which doesn't make any sense. I'll see any anyone else local can get me some.
It almost looks like the DC offset of the other board moves with the pot too, which means it might be the change of impedance... It's just a lot more noticeable on the right board because the initial offset was much higher then the left. Meaning my only way of fixing this is finding a part that should be quite common, but for some reason is not so. AMB outlined this guide here to eliminate the DC offset using R10/R14:
AMB Laboratories DIY Audio • View topic - large DC offset in left channel
When he says turn it to minimum, the pot should be at it's maximum resistance. AMB in his guide states to use a 10K pot, but the schematic shows a 100K pot, I almost think it should be safer to use a 100K pot. I don't quite understand what he means by "Turn one of the two trimpots to change the offset toward 0mV, a little at a time, and simultaneously turn the other trimpot in the opposite direction.". If the pots already at it's maximum resistance, then what's the point?
I'll also attempt to find some moderate sized TO-220 heatsink that i can use for the output transistors. Something like this:
Digi-Key - 294-1042-ND (Manufacturer - PB1-36CB)
or this:
Digi-Key - 294-1080-ND (Manufacturer - 7-340-2PP-BA)
I think i can partially blame it on thermal run away, since the offset will start at something like ~-1.0mV and as the amp warms up, it'll end up at something like ~-25.0mV and the transistors will be about ~85c.
I'll do some more voltage measurements later dealing with the Jfets, I need a break for now.
Thanks a lot!
Yes the DC SERVO has a very limited 'lock' range, that is, it is able to correct little deviations of DC OFFSET because the LEDs are quite 'insensible' to current changes in the bias network of input CSS. So you must achieve a close to 0 V DC OFFSET before re-connecting both servo integrators. To obtain a more 'wider' lock range, you should place a small resistor (22 - 68 OHM) in series between the LEDs and +/-Vs, but as AMB explained first an initial bias close to 0 DC OFFSET must be done.
Cheers
Arturo
Cheers
Arturo
Ok, So let my get this straight having more voltage forward on a LED means there's more current flowing through it? If that's the case I should put a 22-68ohm resistor to limit the current a little bit more so i can have a bit more freedom when I adjust R10 and R14.
Now for a proposal, since i have some 5K multiturn trim pots available, I think, I could replace R11 and R15 with ~1K resistors. Would this be a liable option?
I'll post measurements of the voltage drop across the output resistors tomorrow.
Hi, I think you are complicating matters.Sinple voltage checking will
lead you to the bad part.OK measure Q1,2 emiters with respect to
ground and the collectors too and compare with good channel.Same
goes with driver transistors Q3,4.If you can use the lab scope,pump
in a 1 volt sinewave and look at the top and bottom half output of the
jfet,that should give you the condition of the jfets.
PS the above is assuming you have zeroed the offset to minimum.
singa
lead you to the bad part.OK measure Q1,2 emiters with respect to
ground and the collectors too and compare with good channel.Same
goes with driver transistors Q3,4.If you can use the lab scope,pump
in a 1 volt sinewave and look at the top and bottom half output of the
jfet,that should give you the condition of the jfets.
PS the above is assuming you have zeroed the offset to minimum.
singa
My post above is dealing with a possible way of removing the offset since I dont have any easy to sourced 100k trimpots. I'll see what I can find tomorrow.
OK, So any of the horrible previous problems have seemed to go away. Which just leaves me with DCservos that cause a elevated DC offset ~+/-10.0mV. So far both channels DC offsets have been trimmed to always stay below 1.0mV. I achieved this by using two 5K trim pots and two 680ohm resistors. I've done this only with the right board
I took a couple of measurements:
Offset influenced by pot (Measurements taken without DC servo)
Right Max: 3.6mV
Right min: 1.5mV
Left Max: 1.7 (~5 minutes of warm up)
Left min: ~1.1mV
Voltage drops across 24ohm output resistors
Average left positive: 0.971V
Average left negative: 0.972V
Average right positive: 0.957V
Average right negative: 0.955V
Votages at the input of the amp (keep in mind there's a ~600ohm resistor at the input):
Left: ~0.4mV
Right: ~0.1mV
I don't have definite measurements with DCservo, but the left channel was somewhere ~8.0mV and the right channel was ~3.0mV
Anything fishy? I'll try some functions with the scope when i have access to it.
I took a couple of measurements:
Offset influenced by pot (Measurements taken without DC servo)
Right Max: 3.6mV
Right min: 1.5mV
Left Max: 1.7 (~5 minutes of warm up)
Left min: ~1.1mV
Voltage drops across 24ohm output resistors
Average left positive: 0.971V
Average left negative: 0.972V
Average right positive: 0.957V
Average right negative: 0.955V
Votages at the input of the amp (keep in mind there's a ~600ohm resistor at the input):
Left: ~0.4mV
Right: ~0.1mV
I don't have definite measurements with DCservo, but the left channel was somewhere ~8.0mV and the right channel was ~3.0mV
Anything fishy? I'll try some functions with the scope when i have access to it.
This is true, I found the DCservo to be quite random at some times. It can make the offset worse if the offset is initially anything above 1.0mV, which is exactly what's happening in the left channel, it's about 13.0mV on average. when i have a chance, I'll defiantly do what i did to the right channel. The right channel's Max offset is ~1.5mV, nice and safe!
I don't know why the voltage drop across the output resistors is so large when the guide says that it should only be about ~38.0mV.
I had a chance to listen to the Dynalo with my grados, I first noticed the soundstage, simply massive! It really brings out the best in the grados, I might up the gain to 5x or 8x... the 33pF compensation capasitor might overcompensate, is this bad?
I am assuming that without servo the max. offset are stable within the limits reported by you.
First note that measuring a 'disconnected' servo is not meaningful, most probably it will swing randomly +/-V depending on output offset. To verify that the servo(s) are working correctly the measurements must be done with the servos 'connected'. Another issue is that when the RC time constant is too little (it falls inside the low end of audio BW) it may induce 'motorboating' instability (a very low freq. oscillation). Resolder the servo circuit to prevent from a 'cold solder' and if that doesn't work, I suggest the integrator cap. replacement, and then the servo chip.
Cheers
Arturo
First note that measuring a 'disconnected' servo is not meaningful, most probably it will swing randomly +/-V depending on output offset. To verify that the servo(s) are working correctly the measurements must be done with the servos 'connected'. Another issue is that when the RC time constant is too little (it falls inside the low end of audio BW) it may induce 'motorboating' instability (a very low freq. oscillation). Resolder the servo circuit to prevent from a 'cold solder' and if that doesn't work, I suggest the integrator cap. replacement, and then the servo chip.
Cheers
Arturo
You have over-biased the output stage with R19=215ohm...bypass it with a piece of wire for a moment and measure output bias again.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.