Tape recorder project

[EmilFrid]

Update!
Last night, I finished most of the work on the controller.
I cut the top plate by hand from the bottom of a broken Mackie D8B Mixer (yeah, electricity prices are killing the wallet) and painted it in a color that compliments the rest of the machine. I also repurposed an old fashioned computer keyboard for safe/ready buttons. Cut the section of number 1-4 out, added wooden sides and screwed it into the chassis on top of 4 small momentary buttons. The far left red button is for edit mode, and the pot + switch is for variable speed.

I managed the arduino situation and got a working counter which i will put either on the controller or the meter bridge which will also house safe/ready lights.

Speaking of which, safe/ready circuit done as well. There was space left on the uno to make such a circuit, so just for kicks I abandoned my old circuit and went with arduino.

I just need to improve debouncing of the switches a bit.

 

[EmilFrid]

I've now made a safe/ready + counter prototype with the uno, which works great.
4 of the analog pins on the uno feed the optocouplers that activates the transistors and hence the relays. This controls the record status of any of the channels, through momentary switches connected to 4 digital pins of the uno.
The 5th relay connects to an arduino pin via a FET which switches on or off depending on high/low status of the pin. This either makes or breaks the global 12 volts supplied to the safe/ready switching circuit, making it so that you can select channels without them going into record before you've pressed play+rec, and vice-versa which makes punch-ins possible. The play and rec button are connected to digital pins 7 and 8, and only when pressed together the output pin (the one feeding the FET) will go high and activate the power switching relay.

The display (not pictured) is fed with information obtained from an encoder located inside one of the rollers in the tape path, through the arduino. It now counts in seconds at 15ips. Analog pin 0 resets the counter to zero.

What I'll do know is make a nicer looking circuit board. Then I'll be able to work more on the audio cards.

 

[EmilFrid]

Slowly getting a revised playback card together. The circuit begins with a FET going into an op amp providing IEC equalization (with a pot for calibration) and then two more op amp stages, the last one serving as a buffer. Level adjustable via another pot. And as you see, I've added two lc filters as well. One at the input (to ground), and one before the buffer (in series).

It's not finished, but this is where I'm at now
A very simple core design, but I think it'll be good after som experimentation. I should probably add another calibration parameter, focusing on the head bump.

 

[EmilFrid]

The finished control card for the safe/ready, record functions and digital tape counter. The tape counter works fine if I fast wind to the end of the tape and back to zero perfectly, but if I use it in play mode and return to zero, the counter will still have a significant amount of time left. So I have to troubleshoot this. Anyone has any idea why this is? I was convinced the high speed winding would be the difficult part...

I added an AND-gate to make sure it only goes in to record mode with play + rec pressed down. They first go to one inverter each to make two HIGHS, trough the AND-gate and then into the arduino uno which controls the rest of the record circuitry. Stop, rewind and fast forward resets the record circuitry.

I'm now working on the oscillator. The original fostex oscillator is dead, and I want a module rather than a built-in oscillator, so I'll make one from scratch instead. I'm allergic to stuff built like a monolith rather than a system of cassettes.


Edit: fixed it, just cleaned up the code a bit and added a pull up resistor on the tm1637 data pin. It's now more accurate than the fostex g16 originally was.

 

[EmilFrid]

Oscillator is finished, 150khz carefully compared to the oscillator output of my Teac 85-16. I will post pics next time I'm in the workshop. The oscillator unit is housed together with the audio PSU (regulated from oscillator PSU) in a 1U rack unit. LED:s indicating the state of all 4 rails.
+/-15V and oscillator output via 3 pin XLR. Tape machine has a nice rack in which it'll be horizontally mounted with said unit, NR unit and audio cards mounted underneath.

Ok, so I'm getting greedy with the digital possibilities. Does anyone in here have enough knowledge to at least point me in the right direction of getting an autolocator function into my machine?

 

[ruffrecords]

Oscillator distortion has a direct affect on tape noise. Years ago I had a TEAC A3440. You could erase virgin tape with it and on playback the increase in noise was easily apparent. I could then erase the same tape on a Revox A77 and play it back on the TEAC and the noise was measurably lower. I am not sure how you measure distortion of a 150KHz sine wave.

Cheers

Ian

 

[Analog Pitstop]

I am not sure how you measure distortion of a 150KHz sine wave.

probably trivial: only with a spectrum analyzer (SA) can be done. Skill set to use it: minimal.

The 150KHz needs to resolve about H10, then minimum 1.5MHz bandwidth is required. Modern&cheap USB-SA solutions do this.
How cheap? Depends on the answer "what to measure" question. You need to find out what is the minimum THD you can accept and which still have relevance on noise.

 

[Analog Pitstop]

You need to find out what is the minimum THD you can accept and which still have relevance on noise.

good luck with finding here help for these numbers.

 

[cyrano]

Oscillator distortion has a direct affect on tape noise. Years ago I had a TEAC A3440. You could erase virgin tape with it and on playback the increase in noise was easily apparent. I could then erase the same tape on a Revox A77 and play it back on the TEAC and the noise was measurably lower. I am not sure how you measure distortion of a 150KHz sine wave.

Cheers

Ian

Thanks, Ian. I had no idea...

 

[Analog Pitstop]

the more useful (and rare) are the SA which can measure both FM noise and AM noise separately.
Here you probably need only the AM noise of the 150KHz oscillator.

 

[EmilFrid]

@ruffrecords Thanks Ian! I'm aware of this. This far, I've checked it with a scope to confirm a nice, clean sine wave lining up with the oscillator of the 85-16, and since I know the looks of a wave won't give the full picture of its quality, I will invest in a suitable spectrum analyzer when I'm not broke as a joke. Hopefully that'll aid me in finding any hard-to-spot anomalies.

Re:the autolocate function: I actually found an autolocator code on github.
Perhaps I can work with that.

https://github.com/giokincade/autolocator

 

[ruffrecords]

Most cheap digital scopes these days have an FFT function. Most have only 8 bit resolution but with oversampling that should increase to 10 or 12 bits in practice. so you should be able to see a harmonic 70dB down. Funny thing is the Revox oscillator is nothing special but it is measurably quieter..

Cheers

ian

 

[Analog Pitstop]

which leads to the point that software is an equally important factor in choosing a product.
I only have experience with PicoTech in the cheap side of USB instruments. Not the greatest software, but good enough for typical measurements (and border of acceptable/unsuitable for special measurements).
BTW, I have the Picoscope 6246 model (16bits native, 20bits effective with oversampling, ca. 90dB SFDR, +/-20V AC/DC input range and a high quality swept signal source). This is not the cheapest model they offer. I find it a great device for audio anyhow. The software is the same for all their models.

If Ian tells -70dB is okay for Osc harmonics and tape noise, it means that scope/SA's noise floor should be at least 80-90dB relative to its full scale. You can compute the needed bits resolution (dynamic range) based on this.

 

[Analog Pitstop]

if you are good at HW synthesis, choose a >10k€ Teledyne acquisition card with FDK Option as birthday gift. This will allow you to write your own FPGA code on top of their system board FPGA. Then code all special functions for "on-board real time" measurement you wish and write a software application for it.

To my info, the NI's acq boards (and others), while you can do a pretty flexible SW-dev on them, do not offer possibility for User Area in their FPGA (so, no FW-dev with customized user defined functions, and no extra real time on-board processing functions other than what they offer basic as standard).

 

[Analog Pitstop]

don't forget to code with dBu or dBV for voltages, not dBm!

 

[EmilFrid]

@ruffrecords I've been lurking around online looking at some cheapo digital scopes actually. Not sure how cheap is too cheap though. I want it to be decent. I'd probably benefit from it for other projects as well. And thanks for the useful info, @Analog Pitstop. I will look into this.

 

[ruffrecords]

I mentioned -70dB because that would be a good signal to noise ratio for the typical track width on a pro multi-track recorder without noise reduction so it seemed to me that oscillator distortion probably ought to be no more than this.

Cheers

Ian

 

[Analog Pitstop]

I mentioned -70dB because that would be a good signal to noise ratio for the typical track width on a pro multi-track recorder without noise reduction so it seemed to me that oscillator distortion probably ought to be no more than this.

Come on, you know that a link or more detailed explanation is needed for this statement: mapping the SA results to the SNR results obtained with SNR measurement devices for tape recorders is not a trivial matter.
Else, it means no help. On the opposite, it risks only more deep confusions.

Ian, I thought that you will argument with large bandwidth "noise floor" observed in SA: virgin vs. erased, or virgin vs. biased.
Not with the "SNR" of tape recorder, observed with specialized instruments and precise methods (at 333HZ or so)...

 

[Analog Pitstop]

Example, this:


Here, what is the expected SNR if measured on a SNR device, not a SA device?? 70dB, or 90dB? Or something else?

On SA spectrum: the "blob" around the recorded signal complicates the SNR statements relative to Tape machine datasheet.

 

[ruffrecords]

Come on, you know that a link or more detailed explanation is needed for this statement: mapping the SA results to the SNR results obtained with SNR measurement devices for tape recorders is not a trivial matter.

I agree. I have no idea exactly how oscillator distortion maps to virgin vs erased tape noise. But if you want to investigate it, a good starting point would be to use an SA with a per root Hz noise floor at least as good as the SNR of a good pro recorder. If I knew more I would have said so.

Cheers

Ian