microcassette preamplifier hiss

[Audioguru]

The motor makes a rotating magnetic field that is picked up by the tape playback head because they are too close together.

 

[Easy peasy]

Or, if the motor runs off the same DC supply as the amplifying electronics - this can be a cause of hum pick - up

 

[FvM]

To check if the hum is picked up by the read head or coupled to the amplifier, you need to replace the head with a short. Leaving the amplifier input open cancels the first stage gain and hence reduces hum coupled through the amplifier supply or ground loops.

 

[neazoi]

To check if the hum is picked up by the read head or coupled to the amplifier, you need to replace the head with a short. Leaving the amplifier input open cancels the first stage gain and hence reduces hum coupled through the amplifier supply or ground loops.

I did that now. I shorted the head + to the GND. The same hum is there. The hum is of low volume and not heard during music/voice.
I previously had hum caused by bad DC filtering and when I placed a 2200uF electrolytic at pin 3 of the first opamp (the hum was eliminated. But this hum I describe now is varied as I vary the motor speed, not constant.
I also put a very large electrolytic across the motor contacts. No difference.
The preamp is sensitive to microphonics. Tapping the preamp pcb, this tapping is heard on the speakers. So I held temporarily the motor-player with my hand on the air, so that it does not come in contact with the desk where the preamp sits on. Again no difference.
The only thing I haven't tried yet is a star ground, but the whole circuit sits on the air above a common 5x5cm PCB ground so the distances are small. It is something with the motor/head assembly I suspect. But it mau be normal behaviour. But why would this be normal? Interested to figure it out.

As these microcassette recorders weren't designed for hi-fi, maybe it is an acceptable design issue. Maybe the belt couples vibrations to the head, or maybe the motor indices some EM field to the head as audioguru said?

 

[betwixt]

If I have to be critical...

Why on Earth are the amplifiers fed from an unspecified 13.5V supply but the motor fed from a chain of three regulators. Doesn't make the slightest sense to me.

It is unusual for motors in cassette mechanisms to run from 1.42V but regardless of that, a simple single stage regulator, ideally a switching type will work perfectly well and will not introduce any audible noise. That would eliminate 3 ICs and hopefully kill the motor noise in one step.

DC erase is not a good solution. It works but the erase head will to some extent record noise from the supply. it is basically the same as the record/playback head but with a wider gap and hence no ability to impose high frequency fields to the tape. It will still record low frequencies though, and running it from the incoming supply might add to the recording noise instead of erasing it.

In record mode, there is no bias to the record/playback head. This WILL introduce distortion, primarily at the second harmonic of the audio. There will be hysteresis to overcome as the audio signal reverses polarity. It needs bias, ideally AC bias or the same problem as the erase head will occur.

That 100nF capacitor between the stages really should have a resistor in series with it. The second stage in particular will be very prone to instability without it. The gain is set by the ratio of feedback path impedance and the source impedance. As the source is the output of the first stage it will be very low so the second stage will be running at very high gain. The 100nF capacitor is essentially the only thing controlling the gain and given that it is reactive, the gain will be higher as its reactance drops, in other words at higher frequencies. This is exactly the opposite of what you are trying to achieve.

Brian.

[edit] I should point out that in a previous life I worked as design engineer for a well respected studio recording equipment manufacturer - on the design of tape recorders.

 

[neazoi]

Why on Earth are the amplifiers fed from an unspecified 13.5V supply but the motor fed from a chain of three regulators.
It is unusual for motors in cassette mechanisms to run from 1.42V but regardless of that, a simple single stage regulator, ideally a switching type will work perfectly well and will not introduce any audible noise. That would eliminate 3 ICs and hopefully kill the motor noise in one step.

The mechanism has come out of a portable microcassette recorder. It uses two 1.5v batteries in series. Before I ruined this, I measured the voltages at operation (I just measured DC on the erase head though, not AC bias, if it had any). The motor voltage at 2.4cm/S was 1.42v. The equipment I am powering this now (an FT-301 transceiver) outputs 13.5v. So I have to drop the voltage to 1.42v. Having this tape circuit inside the same enclosure with other RF equipment may induce RF signals from harmonics of switching regulators, that's why I use linear ones. Although to avoid a heatsink I used 3 in them in series. It works. That is the story.

DC erase is not a good solution. It works but the erase head will to some extent record noise from the supply. it is basically the same as the record/playback head but with a wider gap and hence no ability to impose high frequency fields to the tape. It will still record low frequencies though, and running it from the incoming supply might add to the recording noise instead of erasing it.

Maybe another big capacitor connected before the 1k of the erase head to the groung, and a parallel resistor to it (for discharging) will solve any noise issues from the PSU if they appear (I have not tested the record function yet). Remember, this is to record voice only 3KHz max so no high fidelity needed here. And the erase coil and the play/record coil are inside the same head! Something I haven't seen before...

In record mode, there is no bias to the record/playback head. This WILL introduce distortion, primarily at the second harmonic of the audio. There will be hysteresis to overcome as the audio signal reverses polarity. It needs bias, ideally AC bias or the same problem as the erase head will occur.

If I find the sound to be unacceptable I will try to bias it. Any simple circuits to do so? A single transistor oscillator etc.

That 100nF capacitor between the stages really should have a resistor in series with it. The second stage in particular will be very prone to instability without it. The gain is set by the ratio of feedback path impedance and the source impedance. As the source is the output of the first stage it will be very low so the second stage will be running at very high gain. The 100nF capacitor is essentially the only thing controlling the gain and given that it is reactive, the gain will be higher as its reactance drops, in other words at higher frequencies. This is exactly the opposite of what you are trying to achieve.

I see. Despite I hear no problems as it is (tested), I will try to put a resistor there, although I do not want it to be high value to drop the gain much. That would require a third opamp, like in the original schematic...

 

[betwixt]

Maybe another big capacitor connected before the 1k of the erase head to the groung, and a parallel resistor to it (for discharging) will solve any noise issues from the PSU if they appear

The head will have a DC resistance of only a few Ohms so there is no need to add a discharge resistor. It isn't the ideal solution but if you want to stay with DC erase, place the capacitor directly across the head, something like 100uF should be enough.

If I find the sound to be unacceptable I will try to bias it. Any simple circuits to do so? A single transistor oscillator etc.

AC bias is by far the best for recording bias and for erasure. You need a fairly clean sine wave at a frequency around 100KHz for best results. Start with quite a high voltage >20V and drop it through a resistor to the top of the record head. The exact voltage need will vary from one head to another because it depends on the head inductance. The usual trick is to use something like 1K with a 10K pot in series. Record a high frequency audio tone (~5KHz) and play it back, adjust the pot to get the highest playback level so you might have to make several short recordings. When you have the best record level, adjust the pot to increase the bias current until the level drops by about 2dB. Most bias oscillators use a single transistor and a small ferrite transformer.
If you really want to keep it simple but with inferior results, you can use DC recording bias. Limit the current through the head to no more than 1mA through two resistors (measure the head resistance and use Ohms Law) and from their junction with each other, add a capacitor to the head ground. The capacitor is to filter supply noise out of the bias current, something like 100uF will be adequate. Again, there will be a discharge path through the resistors already so no need to add anything else.

I see. Despite I hear no problems as it is (tested), I will try to put a resistor there, although I do not want it to be high value to drop the gain much. That would require a third opamp, like in the original schematic...

Adding a resistor, provided it isn't too high in value might actually increase the gain. It would be worth your while looking at the output on your oscilloscope to check for instability.

I appreciate voice quality doesn't have to be great to be recognizable but if you start trying to record data you will soon find the inadequacies of the recording are a problem.

Brian.