Setting bias on Tandberg 12

[Eichberger]

I followed your suggestion TIMTAPE and the plots are attached.

Unfortunately the noise from the tape is only slightly greater than the noise from the circuits. I can't really discern much from this.

[TIMTAPE]

For the test to be valid for the slower speeds, we'd need to compare plots of the circuit noise at the relevent speed so that the playback EQ's are the same. So there would need to be six graphs altogether, paired for tape speed.

Even so, looking at 2 (tape hiss at 7.5ips) relative to 1 (circuit noise at 7.5ips EQ) I see a significant increase in noise below 10kHz but not above 10kHz, or actually a little higher. (The effect is easiest to see by repeatedly toggling back and forth between plot 1 and 2.) This tallies well with your reported concern about a sharp drop in response above 10kHz in record/playback.

This is what we'd expect if the head/tape contact is not optimal, ie: there is some spacing loss, or the head gap has started to open out. With known good tape to head contact at 7.5ips I'd expect the increase in noise to extend up to 20kHz. There should be no shelving around 10kHz.
This doesnt prove conclusively it's a spacing loss problem but it's consistent with it.

[Eichberger]

Thanks for taking the time to look at those TIMTAPE.

The circuit noise plots at lower speed looked identical to the 7.5 ips version. I will have another go at it later to see if I can discern the peaks from the eq circuit.

Unfortunately another problem has shown up. The right channel is sporadically dropping right down and distorting. People on Tapeheads forum have suggested the mode switches or volume pots as a possible source but I'm not sure about this. Normally a crackley pot or switch jumps in and out at high speed as you turn it but with this problem it just drops out and stays that way. The I power down and on again the channel briefly comes in but then drops down again. If I leave it a day or so it's fine again. A t first I thought something might be overheating but then it did it cold.

[TIMTAPE]

The circuit noise plots at lower speed looked identical to the 7.5 ips version. I will have another go at it later to see if I can discern the peaks from the eq circuit.

Something wrong there. Even just listening to play with no tape playing you should be able to hear the 6.5kHz noise peak at 1.875ips speed, and should hear this noise change on switching to 3.75ips. Of course the listening volume needs to up high, and turning the treble to max also helps.

[TIMTAPE]

Originally Posted by Eichberger

Unfortunately another problem has shown up. The right channel is sporadically dropping right down and distorting. People on Tapeheads forum have suggested the mode switches or volume pots as a possible source but I'm not sure about this. Normally a crackley pot or switch jumps in and out at high speed as you turn it but with this problem it just drops out and stays that way. The I power down and on again the channel briefly comes in but then drops down again. If I leave it a day or so it's fine again. A t first I thought something might be overheating but then it did it cold.

Possibly the "record/play/amp" switches as per the knobs front centre. Almost certainly some internal maintenance needed. Ah the joys of vintage equipment!

[ricard]

Originally Posted by Eichberger

The circuit noise plots at lower speed looked identical to the 7.5 ips version. I will have another go at it later to see if I can discern the peaks from the eq circuit.

I wonder if there could be something wrong with the linkage from the speed selector mechanism to the EQ switch on the amplifier PCB. I believe there is a brass or copper tube in which a thick wire runs to transfer the knob movement to the switch. One source of problems could be the actual speed switching mechanism. The core of the mechanism is a rather large plastic or die-cast metal piece (can't remember which in the '12, probably plastic) which has the cam for selecting the speed, as well as a part that goes through the chassis and operates the EQ switch. It is not unusual for the piece to develop cracks so that the EQ switch does not operate properly.

Unfortunately another problem has shown up. The right channel is sporadically dropping right down and distorting. People on Tapeheads forum have suggested the mode switches or volume pots as a possible source but I'm not sure about this. Normally a crackley pot or switch jumps in and out at high speed as you turn it but with this problem it just drops out and stays that way. The I power down and on again the channel briefly comes in but then drops down again. If I leave it a day or so it's fine again. A t first I thought something might be overheating but then it did it cold.

I wonder if it could be the onset of 'tin whiskers' in one of the transistors. There are other threads on this in the forum. Basically the inside of tinned metal encapsulated transistors can grow whiskers inwards over time, eventually making contact with the crystal, shorting one or another of the electrodes to the case. The 'it works a bit then stops' is typical of this. Powering on the machine after it has been powered off for a while presumably can cause just enough current to jar the end of the whisker for a while. I had a Philips EL3551 which exhibited this phenomenon. And I've lost count of the number of 12's and 1200X's that have passed through my hands which have been silent on one or both output channels due to tin whisker shorts in the driver transistors on the output amplifier boards.

[Eichberger]

Thanks again TIMTAPE and ricard for taking the time to help me with this.

I wonder if it could be the onset of 'tin whiskers' in one of the transistors. There are other threads on this in the forum. Basically the inside of tinned metal encapsulated transistors can grow whiskers inwards over time, eventually making contact with the crystal, shorting one or another of the electrodes to the case. The 'it works a bit then stops' is typical of this. Powering on the machine after it has been powered off for a while presumably can cause just enough current to jar the end of the whisker for a while. I had a Philips EL3551 which exhibited this phenomenon. And I've lost count of the number of 12's and 1200X's that have passed through my hands which have been silent on one or both output channels due to tin whisker shorts in the driver transistors on the output amplifier boards.

Yes, I think it is one of the transistors in the record pre-amp circuit as it does not effect tape playback. So that is Q112 (BC109) Q113 (BC114) and Q113 (BC113). They all look available to buy new. Is there any particular type/make you would recommend? Should I just replace them all or is there a way to identify the failing one without an oscilloscope?

Originally Posted by TIMTAPE

Something wrong there. Even just listening to play with no tape playing you should be able to hear the 6.5kHz noise peak at 1.875ips speed, and should hear this noise change on switching to 3.75ips. Of course the listening volume needs to up high, and turning the treble to max also helps.

I've been a total dunce with this. I did all the testing using an old Glensound mixer which is very very noisy. So those plots were mostly showing its noise!

Am going to do it all again later and also have a look at the speed linkage that ricard mentioned.

[ricard]

Originally Posted by Eichberger

Yes, I think it is one of the transistors in the record pre-amp circuit as it does not effect tape playback. So that is Q112 (BC109) Q113 (BC114) and Q113 (BC113). They all look available to buy new. Is there any particular type/make you would recommend? Should I just replace them all or is there a way to identify the failing one without an oscilloscope?

Sorry, I should have checked the schematics before venturing my comment; to my knowledge, it's only Germanium transistors such as AF126 and AC187 to name a few that have been shown to have the tin whiskers problem. I'm treading deep here, but I think it may be because the internal design differs: Germanium transistors of that era are of the diffusion type, with the transistor essentially in the middle of the package, whereas Silicon transistors tend to be of the planar type which in turn means that the crystal itself is at the bottom of the package, lying flat. Could be that the inside of the can is not tinned either, eliminating the problem completely.

(Of the ones mentioned, only BC109 is in a metal package, and in this design could be replaced by BC549 most certainly).

Anyway, with that transistor lineup I'd rather think it's something else. A 'cold solder' joint or even a bad connector would be my suggestions.
It could also be the mode switch, but then it would help (immediately) to wiggle it back and forth. (Although the surge of current as a result of the switching could 'cure' a bad solder joint for a while, too).

(When it comes to the mode switch there are essentially two versions of the electronics, the breakpoint being serial number 220 9550 . Machines below this serial number tend to have scratchy mode switches which are hard to get working properly (judicious cleaning will help for a while), whereas serial numbers above are fine; might need the switch to be operated from time to time to avoid oxide buildup but that is all. I believe the reason is that originally, the switches pass DC current, which coupled with unfavorable switch materials cause corrosion or oxide buildup, whereas in the later version, the switches don't have any DC across them.)

[Eichberger]

Originally Posted by ricard

I wonder if there could be something wrong with the linkage from the speed selector mechanism to the EQ switch on the amplifier PCB. I believe there is a brass or copper tube in which a thick wire runs to transfer the knob movement to the switch. One source of problems could be the actual speed switching mechanism. The core of the mechanism is a rather large plastic or die-cast metal piece (can't remember which in the '12, probably plastic) which has the cam for selecting the speed, as well as a part that goes through the chassis and operates the EQ switch. It is not unusual for the piece to develop cracks so that the EQ switch does not operate properly.

Spot on ricard. The switch was stuck in the 3.75 ips position, hence the sharp peak and then fall from 10kHz. I've set it to 7.5. As I only ever plan to use this speed I don't need to fix the problem.

TIMTAPE, as you predicted, once I moved the switch there was a considerable change in the noise spectrum.

Loving the learning I'm getting from doing this. Will try to figure the channel two drop out issue tomorrow.

[Eichberger]

I have done a new test with white noise. Things are greatly improved now that the EQ switch is in the correct position but there is still a 10 db drop from about 6kHz to 16kHz and then it goes off a cliff. This is with the bias trim at minimum.

[TIMTAPE]

If everything is in good shape, at a certain low level of bias, the 10 to 15kHz signal should reach a maximum and be actually well higher than say 1kHz. Then normally bias is increased until those high frequencies are somewhat erased by the bias signal and we have a flattish response. This is called "overbias" and is the normal operating condition.

If at no level of bias can we even achieve a roughly flat response at 7.5ips, then again there is something wrong. Again I would check the head/tape interface.

[Eichberger]

I have looked more closely at the head and have done some adjusting without result. I even tried wedging a small piece of soft foam in to gently push the tape against the head (like the pad on the erase head) but this made no difference to frequency response. From looking at a few pictures on line it doesn't look like the gap has opened up and the worn part of the head is only about 3mm across.

One thing I've noticed is that there is a big difference between the two channels. Will do some more plots to see if I can figure anything out but I'm beginning to run out of ideas!

[TIMTAPE]

It's not too hard to remove the head. The brass band over the top holds the head in place and is secured in slots by two small washers either side of the base of the head. The washers can be carefully prized out and the band squeezed towards the head. Unsolder the wires of course. Much easier to view the head face out of the machine. Relapping if needed can be done with simple, inexpensive materials and careful technique.

[mitajohn]

I suggest you to follow this experiment, first to the Left channel (and then the same to Right channel) :

At the tape preamp C113 (1000pF) via C111 (4.7uF) is connected in parallel to the tape head terminals when in playback mode. This works as a filter to bust the high frequency response of the tape head. Try to add in parallel to C113 a 220 or 330 or 470 or 1000 or 1470pF cap and see what happens to the playback response for each cap addition. Post the results.

[TIMTAPE]

Originally Posted by Eichberger

...One thing I've noticed is that there is a big difference between the two channels...

I think yours is 1/4 track stereo. Often the left channel is down because of head wear and as a result dirt caught in the upper corner of the wear groove which is on the left channel pole piece. Removing the dirt can be difficult because it's it a tight crevice. The right channel is not so affected because it's inboard from the wear groove.

One of the reasons to relap the head face is to eliminate the wear groove where the dirt collects.

[Eichberger]

Hi all. Sorry for the radio silence. Had a big family bereavement so fiddling with ancient electronics was out of the question for a bit sadly!!

I just had the head relapped but that does not seemed to have made any difference.

Having had a bit of time to think about it I am sure that the problem must be something that effects both channels equally. So I'm going to open her up and have a look at two things. Firstly I need to look at the equaliser circuit inductors. I was also wondering if the previous owner might have fiddled with them. The other thing is the play/record/amp switches as they will alter the equalisation circuit settings.

Is there anything else that could effect both channels in the same way?

[TIMTAPE]

Originally Posted by Eichberger

Having had a bit of time to think about it I am sure that the problem must be something that effects both channels equally.

But didnt you say earlier there was a big difference between channels? Has the head relapping solved that problem?

Tim

[Eichberger]

There was a bit of a difference between the channels but by far the most significant problem of high frequency roll off is consistent between the two channels. Both have a very similar curve of 20db loss between 5kHz and 15Khz

I need to get it set up properly with the relapped head first and then I'm going to do a bit more exploring.

[TIMTAPE]

OK, good luck with it.

[Eichberger]

I've done a few investigations. The variable inductors in the equaliser circuit are set correctly. Also the rec/play/amp switches seem to be working ok.

Attached are some plots of the frequency response at the different speeds. To do this I had to manually change the switch which alters the EQ for different speeds. They were all done at minimum bias.

The plots all show a HF roll off but the attenuation is different for different speeds. 7.5ips starts to roll off at about 7kHz, 3.75ips starts to roll off at about 4.5kHz and 1.875ips starts at about 3kHz.

What I really noticed are the peaks in the plots which seem to relate to the equaliser peak frequencies (19kHz for 7.5ips, 11kHz for 3.75ips, 6.5kHz for 1.875ips). The frequencies match those on the plots but the size of this peaks for the 3.75 and 1.875ips speeds are surprising. At 1.875ips there is a 10db rise from 5kHz to 6.5kHz. That's way more gain than the equaliser circuit should be providing. Similarly for 3.75ips there is a rise of 5db between 9.5kHz and 11kHz.

I've realised that doing this at minimum bias might not be ideal so I'll do it again at a mid bias level.

Any thoughts?