The setup.
Hantek 100 Mhz scope.
Probes set on 1X
One probe on output.
Was tracking down opamp over run. On a TDA7294. The output was showing over run and a ringing on a square 1Khz wave.
I found the output of the amp to be pretty stable as in at all freq from 20 to 20Khz the output did not change.
There was just this overshoot and ringing and the plan was to try a few caps to clean it up.
I then connected my free probe to the input of the amp. The only thing different I did this time was I connected the probes while the setup was powered on. Instant magic smoke and a blown chip.
What went wrong. So that I make sure I dont do it again.
The only thing different on this day. Was that I setup my probes as 1x.
On previous days I had used 10x on the probe.
The output signal was fine and putting out a clean wave form showing op-amp over run.
I though hey maybe I should chk the input signal to make sure the input to the amp is clean. When it went pop.
Any insights helpful.
Hantek 100 Mhz scope.
Probes set on 1X
One probe on output.
Was tracking down opamp over run. On a TDA7294. The output was showing over run and a ringing on a square 1Khz wave.
I found the output of the amp to be pretty stable as in at all freq from 20 to 20Khz the output did not change.
There was just this overshoot and ringing and the plan was to try a few caps to clean it up.
I then connected my free probe to the input of the amp. The only thing different I did this time was I connected the probes while the setup was powered on. Instant magic smoke and a blown chip.
What went wrong. So that I make sure I dont do it again.
The only thing different on this day. Was that I setup my probes as 1x.
On previous days I had used 10x on the probe.
The output signal was fine and putting out a clean wave form showing op-amp over run.
I though hey maybe I should chk the input signal to make sure the input to the amp is clean. When it went pop.
Any insights helpful.
If you scoped it at the chip pins there is a risk of touching the two input pins together.
Touching those pins together risks making it oscillate at several Mhz at high power.
The probe needs to go the other side of the coupling cap at least.
Also keep the scope earth away from the speaker output.
Touching those pins together risks making it oscillate at several Mhz at high power.
The probe needs to go the other side of the coupling cap at least.
Also keep the scope earth away from the speaker output.
I had one channel on the speaker output i.e. disconnected the speaker and connected the wires to the scope.
Then I had my function generator (Phone Jack). With 3 wires. I connected the alligator to the earth point of the trim pot. And the probe to center point of the trimpot. I know my function gen has a peak to peak off only about 1 volt.
Then I had my function generator (Phone Jack). With 3 wires. I connected the alligator to the earth point of the trim pot. And the probe to center point of the trimpot. I know my function gen has a peak to peak off only about 1 volt.
Hi chinoy,
Good advice above.
You mention probing at a trim pot. Would you post your schematic so we can understand what circuit nodes you're probing?
Another possibility is that switching to using x1 probe provoked oscillation; x10 probes will present about 10 to 12 pF loading but x1 probes present a much larger load capacitance and can sometime provoke amplifier instability/oscillation. I try to avoid x1 probes unless absoletely necessary for added sensitivity; sometime I probe with 220 ohm in series with x1 probes to avoid oscillation.
Be cautious about testing with high frequency, high amplitude square waves. They can smoke the series R/C Zobel loads that are often present on amp outputs.
Good advice above.
You mention probing at a trim pot. Would you post your schematic so we can understand what circuit nodes you're probing?
Another possibility is that switching to using x1 probe provoked oscillation; x10 probes will present about 10 to 12 pF loading but x1 probes present a much larger load capacitance and can sometime provoke amplifier instability/oscillation. I try to avoid x1 probes unless absoletely necessary for added sensitivity; sometime I probe with 220 ohm in series with x1 probes to avoid oscillation.
Be cautious about testing with high frequency, high amplitude square waves. They can smoke the series R/C Zobel loads that are often present on amp outputs.
When you measure clock signals it's better to use an active probe or DIY probe in a configuration that doesn't change the impedance or capacitive loading of the clock source.
I blew up the output section of a CCHS 24.576MHz oscillator by measuring it with a high impedance passive scope probe that has a capacitive load of 15pF, the CCHS has a maximum of 7pF and needs 50R. Reflections from higher impedance can result in a large voltage spike over the working voltage, additionally current draw for the higher capacitance can also damage.
If that clock source was controlling the amp, that would lead the amp freeze at the current clock point resulting in DC which may then cause the amp to go above it's Safe Operating Area (SOA) for some of the components in side (junction temps etc) and POP!
I blew up the output section of a CCHS 24.576MHz oscillator by measuring it with a high impedance passive scope probe that has a capacitive load of 15pF, the CCHS has a maximum of 7pF and needs 50R. Reflections from higher impedance can result in a large voltage spike over the working voltage, additionally current draw for the higher capacitance can also damage.
If that clock source was controlling the amp, that would lead the amp freeze at the current clock point resulting in DC which may then cause the amp to go above it's Safe Operating Area (SOA) for some of the components in side (junction temps etc) and POP!
Maybe this will help
https://audiokarma.org/forums/index.php?threads/basic-amplifier-testing-with-an-oscilloscope.866841/
“Not my neck” you should always have a dummy load, further TDA74xx are like cats
https://audiokarma.org/forums/index.php?threads/basic-amplifier-testing-with-an-oscilloscope.866841/
“Not my neck” you should always have a dummy load, further TDA74xx are like cats
sorry im not good at making schematics. It was phone. Phone has a 3.5mm audio jack. Audio jack goes to 100K trim pot. Output from trim pot goes to amp. Only one channel popped. The speaker did make a loud noise just before it poped.
Its ok I have a replacement chip on the way. But its not a made in Singapore so this will be fund one chanell will be regular TDA7294 and the other will be made in Singapore stamped chip. Lets see if they both play together of id there is any diff. If there is a diff I can change both to 7294 this time I ordered a few.
The made in Singapore matched output of the original chip in terms of peak output.
But there was some op amp over shoot going on when checking square waves. I got a very distorted output in terms of it was square with over shoot and ringing. This data was collected before I tried to tap the output of the phone jack i.e. I was only probing the output at that time.
Its ok I have a replacement chip on the way. But its not a made in Singapore so this will be fund one chanell will be regular TDA7294 and the other will be made in Singapore stamped chip. Lets see if they both play together of id there is any diff. If there is a diff I can change both to 7294 this time I ordered a few.
The made in Singapore matched output of the original chip in terms of peak output.
But there was some op amp over shoot going on when checking square waves. I got a very distorted output in terms of it was square with over shoot and ringing. This data was collected before I tried to tap the output of the phone jack i.e. I was only probing the output at that time.
The new chip is in.
Both channels working.
Back to testing.
And back to confusion.
With no in put hooked up. Only the amp powered on.
Scope connected to output
Logic trying to see if there is any native noise being introduced by the power supply I am detecting an output off about 80 mv.
Im trying to understand what frequency. But if I change the voltage setting on the scope i.e. 100mv / 50 mv / 20 mv as I go lower on the scale the freq of the noise changes i.e. the lower the setting of the voltage scale the higher the freq reading I am getting i.e. at 100 mv I get a freq off 30-150 Hz. If I go down to 20mv scale. The Freq reported by the measure screen shows Mhz.
Is this normal ?. Why should the freq change based on voltage scale ?. Sorry for the noob question. This is a new scope Im trying to figure out. very different from my fluke 105B.
The power section of this PCB is very simple just a 4 pin Rectifier and 4 large caps. I have added some 100 nf between the legs of + and ground and - and ground. On the TDA 7294. Somebody recommended adding a 2uf between the positive and negative but Im not sure that is such a bright idea. The plan was to add some low value NF or PF value caps on the power section to reduce noise. Need to go read the Xmas design post and a few others to see how they filtered the power line.
Both channels working.
Back to testing.
And back to confusion.
With no in put hooked up. Only the amp powered on.
Scope connected to output
Logic trying to see if there is any native noise being introduced by the power supply I am detecting an output off about 80 mv.
Im trying to understand what frequency. But if I change the voltage setting on the scope i.e. 100mv / 50 mv / 20 mv as I go lower on the scale the freq of the noise changes i.e. the lower the setting of the voltage scale the higher the freq reading I am getting i.e. at 100 mv I get a freq off 30-150 Hz. If I go down to 20mv scale. The Freq reported by the measure screen shows Mhz.
Is this normal ?. Why should the freq change based on voltage scale ?. Sorry for the noob question. This is a new scope Im trying to figure out. very different from my fluke 105B.
The power section of this PCB is very simple just a 4 pin Rectifier and 4 large caps. I have added some 100 nf between the legs of + and ground and - and ground. On the TDA 7294. Somebody recommended adding a 2uf between the positive and negative but Im not sure that is such a bright idea. The plan was to add some low value NF or PF value caps on the power section to reduce noise. Need to go read the Xmas design post and a few others to see how they filtered the power line.
I worked on an irs2092 chip and that would blow if I touched scope probe on a pin while it was turned on.
So turn off, connect scope probe then turn on and it should be fine.
So turn off, connect scope probe then turn on and it should be fine.
When probing something, you have to be extra-careful: it doesn't matter whether it is a scope, multimeter, spectrum analyser, audio analyser, the input + cable capacitance can wreak havoc on otherwise totally stable circuits.
A X10 or active probe is the least harmful, but given the wrong spot, it can always pose problems. You have to use your judgment, but in doubt, I always tie a 1 to 10K resistor to the tip of the probe: it may impair the HF response, but generally it kills any potential instability.
A X1 probe is a kind of worst case: you have the full input capacitance + the cable, and this can be >100pF, which is sufficient to upset many circuits: even a placid 741 would balk if it sees such a capacitance on its inverting input
A X10 or active probe is the least harmful, but given the wrong spot, it can always pose problems. You have to use your judgment, but in doubt, I always tie a 1 to 10K resistor to the tip of the probe: it may impair the HF response, but generally it kills any potential instability.
A X1 probe is a kind of worst case: you have the full input capacitance + the cable, and this can be >100pF, which is sufficient to upset many circuits: even a placid 741 would balk if it sees such a capacitance on its inverting input
Thanks. I have observed the same. If you hook up the probes before powering on. Chances of things going south are far less.
Also shifting to using 10X probes no problems after that.