thanks, yeah i'm beginning to realise that, its unfortunate though the cost of them here and the prohibitive shipping cost for used older units to Australia due to bulk/weight. Ham radio clubs aren't as prevalent here, i've already been on the lookout for a while, but may just have to bite the bullet.
There's lots of learning involved in using it properly, but probably easily outstripped by the time spent debugging circuits without one. its become pretty much unavoidable now that i'm starting to play with my own and others more experimental designs rather than the canned type.
There's lots of learning involved in using it properly, but probably easily outstripped by the time spent debugging circuits without one. its become pretty much unavoidable now that i'm starting to play with my own and others more experimental designs rather than the canned type.
I've just burned up my last salvage 4558 in a unity gain PCB and I'm trying to avoid buying anything. I have RC4560, ST33078, and LM4562 in my parts box. The socket already has 0.1 uf ceramic caps on the power supplies within an inch of the op amp. The PCB has 4560's used as LED drivers, so there must be a reason they didn't use them on the music interface!? So I can't put a 22 pf cap on the feedback resistor, there is no feedback resistor. Any other tricks I can do? Anyway to tell on the datasheet whether it is unity gain stable or not? 4558 is 1v/microsec slew rate, 4560 is 5.5 others are faster. I like the 4560's, they are old TI stock made in USA. Would changing the gain to 5 and putting in a 5:1 resistor divider help? This is the op amp on the input echo socket, the other half drives the driver transistor stack. This is a peavey 1.3k amp so if it oscillates it will burn some serious electric power. I do have a scope but just because it is not oscillating today doesn't mean it won't tomorrow.
Last edited:
indianajo,
difficult to say what has happened in your case. The 4558 is pretty indestructable as long as the usual suspects such as supplies are not exceeded or inputs not overdriven from a low impedance high voltage source...
Would have to see the circuit really. The 4560 has far more output current capability... more than the 4562 and NE5532 in fact and capable of operating on even high voltage rails. So that's probably why they are used as LED drivers.
If the opamp you replaced burned up then I thing something else has gone awry...
difficult to say what has happened in your case. The 4558 is pretty indestructable as long as the usual suspects such as supplies are not exceeded or inputs not overdriven from a low impedance high voltage source...
Would have to see the circuit really. The 4560 has far more output current capability... more than the 4562 and NE5532 in fact and capable of operating on even high voltage rails. So that's probably why they are used as LED drivers.
If the opamp you replaced burned up then I thing something else has gone awry...
Yes, I found the leaky diode that took out the 4558. 1n4148 diode reads okay with a dvm forwards & backwards but puts out 12V on pin 3 of the 4558 from the +15 power supply at 24 ma. (It is an input overvoltage clamp). Probably happened when the 4558 originally exploded when the OT's shorted. Have put in the 4560, so far so good - - - -
This PA amp has been a real training ground. I have learned 4 or 5 things Not to do, and a couple of things to do, trying to fix it.
1 don't plug in pin headers 1 pin off.
2 don't get solder on pin headers
3 just because a dvm will find bad semiconductors, doesn't mean ones that pass the diode test are good.
4. Put a room heater in series with the power transformer before trying the output transistors for the first time.
5. TO3p will substitute for TO3 but you have to cut the corners of the heat sink off
6. Metric TO3 transistors require smaller mount screws than TO3 (#4 instead of #6)
7. 100v transistors cost the same as 40v transistors and don't explode as often
8. 3 amp fast recovery rectifiers will substitute for about anything, but won't fit through the hole for a 1n4148.
9. There are lot of reasons why the idle bias current is wrong. Check it before you quit. Clamp your temp sense diodes on the driver pcb with a diode stack (or 3.3 zener I am using, might work?), because if the cable to the O.T. corrodes the bias current runs away.
There's more, just can't think of it right now. oh yeah, check for oscillation before you are done.
This PA amp has been a real training ground. I have learned 4 or 5 things Not to do, and a couple of things to do, trying to fix it.
1 don't plug in pin headers 1 pin off.
2 don't get solder on pin headers
3 just because a dvm will find bad semiconductors, doesn't mean ones that pass the diode test are good.
4. Put a room heater in series with the power transformer before trying the output transistors for the first time.
5. TO3p will substitute for TO3 but you have to cut the corners of the heat sink off
6. Metric TO3 transistors require smaller mount screws than TO3 (#4 instead of #6)
7. 100v transistors cost the same as 40v transistors and don't explode as often
8. 3 amp fast recovery rectifiers will substitute for about anything, but won't fit through the hole for a 1n4148.
9. There are lot of reasons why the idle bias current is wrong. Check it before you quit. Clamp your temp sense diodes on the driver pcb with a diode stack (or 3.3 zener I am using, might work?), because if the cable to the O.T. corrodes the bias current runs away.
There's more, just can't think of it right now. oh yeah, check for oscillation before you are done.
Last edited:
Your welcome
It's easy to begin... just connect the scope "Y" input to the players audio output socket in the first instance. Play a CD, preferably a test tone and see if you can see any hash on the signal. Adjust the scope controls for a stable display. You won't damage anything by connecting to the audio output socket.
Think of the scope input as like an input on an amplifier. It has an outer ground and an inner signal.
Probes come in all shapes and sizes and complexity right down to just a wire with croc clips
For initial experimenting (if you can't find a probe) is just two lengths of wire. The ground of the audio out on the player goes to the ground on the scope. There will probably be a big ground terminal (screw/nut) right on the front which you can use. The other wire just goes in the centre of the "BNC" type connector on the scopes "Y" input and to the centre of the RCA output on the player.
That will get you started. If you have a probe then you can start looking at the opamp outputs directly.
Probes come in all shapes and sizes and complexity right down to just a wire with croc clips
For initial experimenting (if you can't find a probe) is just two lengths of wire. The ground of the audio out on the player goes to the ground on the scope. There will probably be a big ground terminal (screw/nut) right on the front which you can use. The other wire just goes in the centre of the "BNC" type connector on the scopes "Y" input and to the centre of the RCA output on the player.
That will get you started. If you have a probe then you can start looking at the opamp outputs directly.
hi Mooly, thanks for the infos!
I was wondering, is there any way you could please set up a tutorial that would explain how to check for oscillation and especially how to cure it(decoupling caps?)?
There are many inexpensive USB oscilloscopes available, the cheapest that could the job would a strong asset. If a <$100 unit could do the job, I'd buy it and religiously follow your instructions =)
Thanks for your help!
I was wondering, is there any way you could please set up a tutorial that would explain how to check for oscillation and especially how to cure it(decoupling caps?)?
There are many inexpensive USB oscilloscopes available, the cheapest that could the job would a strong asset. If a <$100 unit could do the job, I'd buy it and religiously follow your instructions =)
Thanks for your help!
Last edited:
Scopes that use a soundcard are digital or discrete sampling. They can "miss" the event between the samples. To avoid "missing" the event the sampling frequency needs to be high. A good guide used to be ten times the roll off frequency of an analogue scope.
An analogue scope monitors the waveform continuously. They are better for debugging.
Try for >=20MHz for catching Audio related events like HF oscillation. FETs can oscillate >>100MHz and a 100MHz analogue might not catch these events. A 1Gsample/second digital scope should see these events. Not sound card territory.
An analogue scope monitors the waveform continuously. They are better for debugging.
Try for >=20MHz for catching Audio related events like HF oscillation. FETs can oscillate >>100MHz and a 100MHz analogue might not catch these events. A 1Gsample/second digital scope should see these events. Not sound card territory.
USB scopes usually have limited bandwidth as Andrew implied. They're OK for audio range checks, but will totally miss oscillations. They're a useful adjunct to a "real" scope, but not a replacement. You want 50MHz or better bandwidth, and if you're working with MOS devices, you'll probably want 100-200MHz. Used scopes from top makers with that kind of bandwidth are not terribly expensive, a couple hundred dollars.
Start here with Tangent's tutorial - I found it very informative & helpful
Working with Cranky Op-Amps
will do, thanks!
well, you can see AD797 oscillating like hell in those links: Asus Essence One Headphone Amp/DAC (CeBIT 2011)
not quite an analog oscilloscope either.
well, you can see AD797 oscillating like hell in those links: Asus Essence One Headphone Amp/DAC (CeBIT 2011)
not quite an analog oscilloscope either.
I think everyone has covered the reasons why a USB scope isn't suitable.
As to a tutorial on checking and curing oscillation, well checking is one thing but there are many pitfalls for the unwary. Even something such as the capacitance of a standard scope probe or wires can actually "cure" or in other cases provoke or modify any oscillation.
A big problem is that each and every case is different and particularly so with DIY builds that often have issues with layout and wiring.
Best advice I can give is as others have said and that is to get yourself an analogue scope with a decent bandwidth. I use a 100Mhz scope but before that I had a 20Mhz one and a 30Mhz one "for the day job".
Thanks for the reply, the person who measured AD797 in my last link told me "My oscilloscope is Velleman PCS500 with HP 100 MHz probes. It is good companion to PCG10 function generator and Pc-Lab 2000 software. It has optically isolated parallel port interface".
It can be found for $150 on ebay
You can isolate USB via ADuM4160, so isolation is not the problem here.
It can be found for $150 on ebay
You can isolate USB via ADuM4160, so isolation is not the problem here.
