Good forums to help fix electronics?

[Bryce]

Hast du die beiden Dioden schon getestet? (because we're in the German section )

Have you tested those two diodes yet?

Bryce.

[tjohnson]

Yeah, I must have created it in the german area somehow.  The diodes test ok with my multimeter, have taken the fets out too and tested and they seem ok with a basic MM test.  The slightest touch on a ground point with the scope or meter can trigger the distortion.  I'm back to my theory that it's feedback on the muting circuit, that the circuit is not muting because the processor isn't putting out high enough.  I was thinking of ways to recreate the signal.  The record out signal is inverted, so it's low when playing and high when recording but the muting circuit is reversed.  Do you think I could use some sort of small logic gate IC to pickup this signal and split it, inverting it for the record mute and passing through for record out?  It must be able to pull it low somehow.  The thing I notice is that at the connector I pickup +5.1v high and -2v low, looking at the circuit diagram it seems to connect to the negative 8v rail via a 10k resistor.

[Bryce]

A single NPN transistor and a resistor is all you need to invert a logic bit.

Bryce.

[tjohnson]

Hi Bryce, I found your old post (here) where you show how to create an inverter, my main concern is that if I pull the pin from the processor then measure the base of the transistors I get about 3v.  From the circuit diagram it looks like this is from the +8v line via R663 (4.7k).  however when I plug in the pin from the rpcessor and measure I get the following at the base 1.2v (high) or about 0.1v (low) so somehow overrides this.  Do you think this would work?

[Bryce]

Hi Bryce, I found your old post (here) where you show how to create an inverter, my main concern is that if I pull the pin from the processor then measure the base of the transistors I get about 3v.  From the circuit diagram it looks like this is from the +8v line via R663 (4.7k).  however when I plug in the pin from the rpcessor and measure I get the following at the base 1.2v (high) or about 0.1v (low) so somehow overrides this.  Do you think this would work?

Wait. Are you measuring the voltage on a pin that's not connected to anything? That's a floating voltage and could be anything. You only measure voltages that are in a circuit, because without a current, there's no knowing where it could be. This is why you also only measure a PSU output when it's connected to a load.

As far as whether the inverter will work: It's a Bipolar transistor not a MOSFET, it switches on current, not on voltage. If the current across Base and Emitter is 30mA or so, the transistor will be fully biased and will output a logic LOW. With anything less than that the transistor switches off and the resistor from the 5V to the collector will pull the line high.

Bryce.

[tjohnson]

As I say this is hurting my head!  I'm pretty sure I measuring in circuit which when the rec mute is plugged in I pickup circa 1.2v when in high state and effectively ground when low.  If I pull the pin out I'm still get voltage and drop across resistors and the mute function seems to work so I assume when the processor goes high it mutes and when it goes low it acts as ground and there is low current flowing to the muting transistor base to switch them on.

It seems to work intermittently so maybe this damn processor has taken a hit on this pin but not completly knocked it out.   I'm nervous of messing around with other parts of the circuit but trying to work out a bodge to get this thing working again.  I was thinking could I use a a NPN with the record mute output connected low side.  If the processor is low the transistor is off and it pulls the circuit to ground but if it's high (whatever is left of high) that would supply the current to switch the transistors on.  Another crazy idea I was having was whether to use an arduino to check the voltage at the pin and then switch a pin high or low if it detects. the voltage rise therefore replacing the processor output.

The other one was that a 15k resistor is used on the main muting line so maybe dropping this resistor to slightly would improve it.

[Bryce]

I still think that the issue is either a PSU or a grounding issue. Especially as you said: "The slightest touch on a ground point with the scope or meter can trigger the distortion." ie: The scope (if it's a real scope) is grounding the otherwise floating ground. This definitely shouldn't happen, so you should be focused on finding out why it does happen.


Bryce.

[tjohnson]

I agree it does appear to could be something else and maybe I'm getting fixated on this muting circuit being the issue. 

Do you have any hints on finding grounding issues?  The only thing I've found on the PSU is that the 12v line is very high measuring approx 16v and 14.5v under load when dubbing but I don't think this is likely to be causing an issue as it's used on the motors and mechs which all seem to be working very well and it's not used on the main PCB board.

TBH I'll probably have to throw in the towel on this one, it would be nice to fix it but I'm spending way to much time on it going round in circles ,it's just nagging me now as I want to know what the issue is!

[Bryce]

Connect the crocodile clip of the scope to the best ground point you can find on the secondary side of the PSU. Then probe every other ground point ypou can find around the audio circuits. You are looking for any points that are at a different potential or display any sort of noise.

You should also measure all the voltage rails with your multimeter set to ACV and see if you get any high (above 50mV) readings. This will give you a good indication of the health of the capacitors.

Bryce.

[tjohnson]

Update on this Bryce, I kind of fixed this.  I couldn't find anything else obviously wrong using the method you suggested so as a last ditch attempt I purchased a £2.50 ATtiny85 micro controller board off eBay, set it up to read the voltage from the main processor and set a pin high when it read over 1v and otherwise low.  Wired it into to the 5v on the main pcb for power, removed the mute line from the socket and connected to to the Tiny85 and then fed the new high low signal from the Tiny85 onto the main pcb and it now works perfectly recording and playing.  So it does appear to have been that muting circuit.  Strange I don't know why the mute line is failing to go fully high and put out 5v there must be a fault inside the processor.  I also managed to fix the original issue I was investigating by replacing a couple of variable resistors.  So it looks likes my tape deck is now useable again, how long it lasts is anyones guess.  Thanks for your help !

[Bryce]

It sounds like the CPU has open-collector outputs and the pull-up resistor associated with that signal has gone open circuit. Is there a resistor anywhere between that line you now control with the ATTiny and to the 5V rail? Is the resistor ok and if so, is there really 5V on the rail at the point the resistor connects?

Bryce.

[tjohnson]

I've attached a picture of this part of the circuit.  There is no connection to the 5v rail in the main circuit.  The record mute signal controls 2 sets of transitors.  2 NPN on the input to the record amp and a PNP which then controls 2 NPN on the output of the record amp.
At the top right of the picture this is connected to the +8v rail via a 4.7k resistor.  There is also 15k resistor prior to the 2 input muting transistors which after the 4.7k resisitor in the circuit.  Both these resistor are ok.

At the moment the tiny85 is replacing the high/low signal that would have come from the processor, I've connected the output to the pin on the connector.  When the processor line was connected in the circuit wasn't muting the majority of the time.  When I removed this pin from the connector the circuit was permanently muted so wouldn't record but played ok.  It was working very intermittently with say 9 times out of 10 not muting.

[Bryce]

Unless there are dry joints on the 4K7 resistor or a crack in the PCB trace, it really could be internal damage in the CPU.

Bryce.

[tjohnson]

Unless there are dry joints on the 4K7 resistor or a crack in the PCB trace, it really could be internal damage in the CPU.

Bryce.

I think it's internal damage on the CPU, I've pretty much reflowed the whole board including that resistor.  I didn't see any cracks but I know they can be tiny so I may have missed them.  I have soaked tested my fix and it seems to work and is stable so I think i'll have to stop there.  I've spent inordinate amount of time on this already!  The deck sounds surprisingly good really, tape was never a great format and considering most people scorn these twin decks it makes good recordings.  I think the control board is a little cheap, it has hard soldered connections with hot glue on them, but the  rest seems well made and being single sided is very easy work work on.