Flux Loop for Tape Machine Calibration

[Gold]

It looks like it is the end of an era. Magnetic Reference Laboratory is closing shop. The obvious solution for tape machine calibration is a flux loop. I know very little about the letter B. What would it take to make a calibrated flux loop for a tape machine?

 

[jeffbro]

Sad, sad news indeed.

This means it's time for us DIY'ers to get serious about fabricating flux loops and variable equalizing amplifiers.

It would be great if we could come up with 3D printable wire loop bobbins that will fit all the different tape widths. I'd offer to do it but I'm slow as molasses with Fusion 360.

To drive the loop an Ampex flux loop amplifier (shown in Section 6 of the ATR-100 manual) might be a good start but I think I would ditch the 30 some component discrete power amp on the output for a simple chip amp. Simplify, simplify.

 

[Gold]

I’m not an ATR user so I’ve never seen the manual. I’ll have to get my hands on one.

 

[EmilFrid]

I think this sounds interesting, but I feel I know too little. I think I understand the fundamentals of how a (relative) frequency measurement via flux loop would go about, but how do you calibrate the overall level? Is it just a matter of setting the desk oscillator to 0db, inject it into the flux loop and expect 0db at the line amplifier end in the case of a correctly calibrated PB amp?

 

[CalavoBob]

There's a pretty good Flux Loop for professional tape recorders article actually titled "Magnetic Reproducer Equalization Accuracy or Flatter is Fatter" written by Peter Butt that appeared in Recording Engineer Reproducer magazine in the October 1978 issue. He describes using a flux loop to check the overall playback response of the head and repro amp combination. The repro head naturally exhibits a 6 dB/octave rising response due to Faraday's law of magnetic induction. The head is followed by an integrator amplifier which has a -6 dB/octave characteristic which should ideally bring the frequency response back to normal. In fact the integrator response is made to level off to flat at some frequency which corresponds to one of the various international recording equalization standards using an R-C network. The break frequency of the R-C network is numerically referred to using the time constant "tau" =RxC in seconds. He points out the obvious advantages of using a flux loop to check playback response since the results aren't affected by problems like azimuth, skewing, gap loss, gap non-linearity, spacing loss, fringing effects, noise, and contour effects.

The loop he describes uses 30 turns of 30 gage enameled wire, in series with a resistor (nearly equal to the source resistance) and mounted on a non-ferrous frame of fiberglass vector board, shaped into a rectangular vertical loop with the active face placed directly on center with the gap running vertically across the string of pole pieces. The ones I have made were done on blocks of styrofoam which I was easily able to cut into shape to fit closely over the front of the head. A cable of RG58A is attached securely, and to prevent the wires from scratching the head surface you finish it off with a thin safety epoxy coating on the front face. Because of the many different shapes of playback head covers out there, you will probably need to make special ones to get intimate placement for each tape recorder you work on.

Here's a short synopsis of his test procedure:
1. Monitor the playback signal from the repro amp output, not the tape deck line amp outs (they can have additional frequency rolloffs).
2. Position the flux loop for maximum output at 400 Hz.
3. Sweep the test frequencies up to find the Repro peak and just beyond, to make sure you are not clipping.
4. Here you'll need to consult the recommended playback curves for your chosen standard (15 ips NAB, 30 ips AES, EIA or CCIR). Note the expected dB level for 10 kHz with respect to 400 Hz, and adjust the Hi Freq eq to this setting. Also set the Low Freq eq if available and recheck the 10 kHz level.
5. Calculate the expected theoretical standards levels from 10 kHz and above at 1 kHz intervals past the response peak, and compare these to your actual flux loop measurements. If you find a rise, a maximum and a fall that deviates from the ideal response, you have identified a playback head-to-cable resonance that can be corrected by paralleling a proper damping resistance to the head-cable combination. This is a trial and error effort but Mr Butt suggests resistances in the range of 100 kohms to 430 kohms have worked on tape decks from a variety of manufacturers.

As far as obtaining the response curves for the different international standards, I have seen them included in a manual that was printed by the old Standard Tape Laboratory company. I'm not sure if Magnetic Reference Labs also has these curves available on their website, but it sounds like something they would offer there.

Flux loops can also be used in reverse, that is to pick up the magnetic flux coming out of a record head and monitored on an oscilloscope. The recording process is actually linear (unlike the playback event that has the rising 6 dB/octave characteristic) but as with tape playback standards, the various standards may include deviations from this linear recording event. Again, checking with published record and playback standards graphs is a good idea here.

 

[gyraf]

Darn, I hope MRL helps us out here by making or at least specifying such a canonized fluxometer

I'd still worry about setting azimuth and skew though

/Jakob E.

 

[Brian Roth]

In the late 1970's, I attended "Ampex school" at their Redwood City headquarters. Among many documents, I got a flux loop paper written by Alastair Heaslett who was one of the three engineers that lead the development of the ATR-100. Lots of technical "meat". Here is my scan of that paper from 20+ years ago when I was first working with my new HP scanner.

 

[Gold]

That’s a good post. Thanks.

The thing I want to know is how a calibrated flux level is determined. You can’t make a 250 nW/M test tape unless you know what 250 nW/M is.

So far everything thing I’ve read about flux loops says you still need a calibration tape for absolute level. How do you make a test tape without knowing what the actual flux is. A chicken and egg thing.

 

[Gold]

In disc recording it is incredibly difficult to determine raw level in cm/sec. The only way to do it with fine accuracy is with an optical measurement system. These optical measurement systems were never commercially available. They were all proprietary in house built systems. I know Ortofon had one. I’d assume Neumann had one although I have no evidence one way or the other.

Do you need a flux meter at a calibrated distance?

 

[Brian Roth]

The whole concept of measuring "NanoWebers per Meter" completely boggles my small brain <g>.

Bri

 

[Brian Roth]

The only thing that comes to my tiny brain is to purchase one or more "fresh" MRL tapes, carefully align a pair of well maintained machine, and create multiple "secondary standard" copies for the studio. We did something similar a year or two ago at the request of a customer of our "Ultra Tape" products. It took some time to super-fine-tune a pair of our ATR-100 machines to get as close to as possible to "a gnat's whisker" with the levels and EQ. Normally we have to allow very small deviations (say +/- 0.25 or 0.50 dB over the audio range) but we fought and fought to narrow that as best as we could for those "secondary standards". We made 10 (?) of those, all from the same reel of blank RTM 900. We skipped over the first few minutes of the blank reel before cutting the copies. Even then, each copy was slightly different from each other....likely due to the variations of the oxide down the length of the tape.

 

[Gold]

I’m going to purchase some spare copies. I only playback. I don’t record so my needs are easy. Most of the time I’m lining up to other tones. I use the MRL to troubleshoot and use if there are no tones on a tape, as a starting point.

I have an unhealthy fascination with reference levels and units of measure.

 

[sodderboy]

I never had MRL's in the rooms unless I was doing periodic maintenance. I used a yearly "WORK TONES" reel made on a verified good machine just like you would stripe tones for a project. That way the WORK reel got beat-up setting up new projects. And maybe erased once, but no one fessed-up. Plus you could use a printed 100 Hz for playback alignment with no fringe effect.
For 95% of 2" work no one will miss having MRL around. Tracks are dumped to DAW anyway. As you say Paul, those who are fascinated will need something more.
For a 2trk it should be easier than 24. Starting with a calibrated machine,
1. couldn't you use a demagger as the first coil attempt? Kinda like rapping into your headphones. It's already designed to position a magnetic field in a specific area. Don't even know if a power amp is necessary.
2. create a DAW mix track of 240kHz bias freq. plus 1KHz, I'm thinking 10:1 as a beginning
3. connect the demag to a sacrificial output amp, I would try a cass deck headphone out or mini guitar amp to start, something driving low impedance
4. position the demag tip at the head, using latex or other protection between the two
5. oh, getting tricky, depending on the machine, how do you un-mute a machine in STOP? Playing a blank tape would complicate things but not make them impossible. A80 is no prob there.
6. increase test track signal level to read 0 Vu on the machine's meters and note the measured signal level.
7. adjust bias/tone ratio and output amp depending on results
8. move on to "Ten kilohertz. Set reproducer gain so that the volume indicators read zero decibels"
9. test for repeatability

An accurate coil will be needed for a 24trk, maybe John French kept all those worn REC heads? Or calibration for each channel. The coil is the tricky part with 24trk. Just like making the full track 2" heads made for MRL.
Mike

 

[RSRecords]

Been wanting to make one for a long time. I have a drawing off one somewhere. Messed around with a few turns around a peace of wood. Worked but was cumbersome. 3d printing a bobbin would be awesome

 

[iomegaman]

Oddly late to the party but almost on time...

I won an auction on a Fostex B16 last week for $102.00 (ended up being "Local pickup only" which was why I won and required a 2 hour trip to Phx to pick it up)...


That same night when I got home I found a local guy on facebook selling a new roll of ATR 456 1/2" never been opened for $80...go figure...I ended up giving him $150 and got 3 rolls of 1/2" and 7 rolls of 2" from the estate of the guy who played keyboards for Kool and the Gang...ended up being $15 a roll for some good stuff and some stuff questionable...

That night I decided to go ahead and purchase the 1/2" MRL tape on reverb for $65.00...might as well...

Then Friday I see that MRL is closing shop...


I'm going through the audio boards one at a time (I need to build and extender card) but right out of the box this machine is working mostly...

Just ordered a remote for it on ebay for $150...oddly enough located 2 hours away...

This idea of Flux loop calibration is right up my alley since I have no way of calibrating the 2" Bouse machine I am still working on...there is no manual, no calibration instructions, nothing...guess work all of it.

I will have to do some math to use the MRL tape since it was recorded at 30 ips and the Fostex is only 15 ips...but I have plenty to do before then.

Gonna dig into this flux loop stuff...makes a certain amount of sense...and obviously there needs to be a way forward for these machines...

 

[terry setter]

You can make a "backup" version of the MRL test tape. I suppose you could copy an MRL tape to another machine, but what I used to do was, just after calibrating my machine, I'd record the same signals at the same levels as are on the original MRL tape. If the machine is calibrated correctly, it will be a very good substitute for the original.

 

[RSRecords]

Problem with a flux loop is also its benefit. It totally removes tape to head contact from the alignment procedure. Great for testing and designing electronics but I don’t know how well a calibration would translate since no tape path is perfect. Flux loop would still lighten the load on an MRL for testing and rough cal.

 

[iomegaman]

A flux loop as far as I can tell won't help with azimuth...I've heard of a method where you record a tone, flip the tape over and play it back to check azimuth...but it seems tedious at best...

Surely there's some way to keep the technology alive...

To be fair azimuth seems like a mechanical issue that really shouldn't come up that often but then there's the randomness that is part of evolution and entropy of everything.

 

[waltzingbear]

There are tables that give the correction factors for wrong speed playing of test tapes on the MRL web site. Also for different eqs.
If you go back far enough in the internet archive you can find the link as it is currently doesn’t have a url posted. Hooray for the way back machine

 

[Brian Roth]

This paper has the conversion tables:

https://www.mrltapes.com/choo&u.pdf
Bri