A800 Noise Floor Testing

[Pat Maki]

With the help of several of the good people in this community, I've managed to bring a somewhat long in the tooth A800 MKIII back into operation and have been doing some test recordings recently. Most caps have been replaced throughout, the heads have been aligned and all meters and cards have been calibrated. What I've noticed in the recordings is that some of the channels have an almost undetectable noise floor, others have a bit more noise, but still low and a couple have, well, an unacceptable noise level.

My current questions to the good folks here:

• What is the scientific way to measure the noise level on the recorded tape? I've found several approaches in the internet, but not clear on this.
• Iis there a published maximum target noise level by Studer? If not, what level do folks shoot for? Been unable to find this...
• What are the key factors the generate too much noise? ie., out of spec components? improper calibration? other? For the couple channels that are pretty noisy, they've been given new caps like the quieter channels as well as the cards being calibrated in the same way. I've completely disassembled and reassembled the machine including refitting/reseating all connectors, so I doubt it's a wiring issue, but who knows....there some many connections in here and all have a chance to glitch.

Many thanks in advance for your thoughtful replies.

Pat Maki

 

[gyraf]

have thoroughly defluxed the heads?

 

[ruffrecords]

There are three sources of noise you need to consider and preferably measure.

1. The noise of the replay amp itself. You can usually measure this in playback mode with the tape paused. THis noise level is usually a good 10dB below virgin tape noise.
2. Tape noise. Measure this by playing some virgin tape. This noise level is usually significantly higher than replay amp noise.
3. Bias oscillator noise. Measure this by recording no signal to a blank tape. This noise level is usually a couple of dB above virgin tape noise but it in a well designed recorder it can actually be lower than virgin tape noise. However, in a poorly designed or faulty recorder it can be much higher and my suspicion is that this is the source of your problem.

Cheers

Ian

 

[Pat Maki]

have thoroughly defluxed the heads?

Actually it's been a few sessions since I've demagnetized. Will do so before continuing. Thanks for the reminder!

 

[Pat Maki]

There are three sources of noise you need to consider and preferably measure.

1. The noise of the replay amp itself. You can usually measure this in playback mode with the tape paused. THis noise level is usually a good 10dB below virgin tape noise.
2. Tape noise. Measure this by playing some virgin tape. This noise level is usually significantly higher than replay amp noise.
3. Bias oscillator noise. Measure this by recording no signal to a blank tape. This noise level is usually a couple of dB above virgin tape noise but it in a well designed recorder it can actually be lower than virgin tape noise. However, in a poorly designed or faulty recorder it can be much higher and my suspicion is that this is the source of your problem.

Cheers

Ian

Thank-you Ian! This is exactly what I was hoping for. A couple clarification questions:

1. Suggestions on most accurate way to read dB's? Ie., checking output voltage and converting it? using VU plugin? etc...I have a scope and some plugins like REW and NAK T-100.
2. Replay amp - Don't have a pause function on this one, so wondering if simply having the channels in Repro mode and the machine on will suffice? If not, possibly running in Play mode without tape loaded?

Br,
Pat

 

[RuudNL]

Running in play without tape will usually work.
I think you only have to 'fool' the light sensitive cell by covering it...
(And maybe simulate tape tension.)

 

[zamproject]

Iis there a published maximum target noise level by Studer? If not, what level do folks shoot for? Been unable to find this...

Studer service manual are usually very exhaustive !!! have a look at it.
S/N is given in section 1 -general- / -technical spec- for various speed, repro and sync, as for 8/16 and 24trks versions
I don't look much but there is probably a level diagram too.

 

[Pat Maki]

Running in play without tape will usually work.
I think you only have to 'fool' the light sensitive cell by covering it...
(And maybe simulate tape tension.)

Thanks for confirming @RuudNL. I'll plan to do this little dance with the machine (-;

 

[Pat Maki]

Studer service manual are usually very exhaustive !!! have a look at it.
S/N is given in section 1 -general- / -technical spec- for various speed, repro and sync, as for 8/16 and 24trks versions
I don't look much but there is probably a level diagram too.

Hi, yes, thank-you for the reminder. I have this manual indexed well in PDF ....indeed very thorough!
Question, for the Record-Repro s/n ratio limit of 66db. Does that mean the noise level on a recorded tape should be no more than 66db above the noise level on a virgin tape....or does it mean something else?

 

[RuudNL]

I suppose the 66 dB is the range between the (weighted?) noise (on a recorded tape with bias only) and a reference level, this could be 320 nWb/m or 510 nWb/m. You need to know both variables (weighted/unweighted and reference level) to obtain a reliable measurement.

 

[zamproject]

Not sure...
there is a statement about ref minus 6dB to operating level, with specific magnetization and tape type

whatever the figure is it's probably for virgin tape and well calibrated recorder strait out from the factory.

@RuudNL , unweighted (NAB std), 510 nWb/m

 

[radardoug]

Buy an old HP 334 distortion meter. Then you can read your signal to noise, and the output from the meter goes to a monitor amp, so you can hear the noise amplified.

 

[Pat Maki]

Buy an old HP 334 distortion meter. Then you can read your signal to noise, and the output from the meter goes to a monitor amp, so you can hear the noise amplified.

Now we're talking! Will take any excuse to acquire cool old gear. These look pretty reasonably priced. To try to get a better feel for this with what I have I recorded 3 things into PT, 1. Standby (machine on in Repro mode but not moving, 2. Virgin tape being played through Repro head 3. Virgin tape recording nothing and playback through Repro head). This is fascinating to look at and to help visually see which channels and modes/cards need to be addressed. I'm going to see if I can track down a noise meter plugin that's sensitive enough to see the dB variations between each mode....these waveforms are enlarged to the max....actual noise is not that bad! (-; Thoughts on this approach?

 

[Pat Maki]

Not sure...
there is a statement about ref minus 6dB to operating level, with specific magnetization and tape type

whatever the figure is it's probably for virgin tape and well calibrated recorder strait out from the factory.

@RuudNL , unweighted (NAB std), 510 nWb/m

Thanks both. Will marinate on this idea.

 

[RuudNL]

In case of doubt, my rule is always: if it sounds good, it is good...

 

[ruffrecords]

The simplest way to measure noise these days is to use a regular USB audio interface with a PC program called REW. This includes a real time spectrum analyser so you can actually see the noise floor and it will give you a direct reading of noise level in dBU. You need to calibrate it to 0dBu but that is easy enough to do.

Cheers

Ian

 

[Pat Maki]

The simplest way to measure noise these days is to use a regular USB audio interface with a PC program called REW. This includes a real time spectrum analyser so you can actually see the noise floor and it will give you a direct reading of noise level in dBU. You need to calibrate it to 0dBu but that is easy enough to do.

Cheers

Ian

Thanks Ian. Glad to hear that as I already have REW. Was using it for other parts of the cal/alignment process. Recording all the tracks to PT with no inputs (in standby, then play, then record) all through the repro head has also been very revealing as well. In solo, track 3 sounds to have the lowest noise and most hifi. This was confirmed visually by the PT session (image in prev. post). I'll start by try to match all tracks to that spec! I'd be happy to get them all there! Also realized that over the weekend, i hadn't update the elcaps in the Hi Freq cards 9-24 and with the wife and daughter out of town, was able to go burn many an hour and got those all updated along with some others. It seems the highest noise is on the channels where I still have some old elcaps in the VU's modules. These will be next. Thank-you again for your invaluable guidance!

 

[tinn]

Calibration Tape?

Sometimes a calibration tape is not available or old and unreliable.
You may check the gain of the repro amps with driving the recording into audible/visible saturation
and check if these distorted repro levels are halfway equal on repro. 400..500 Hz seemed convenient to me. The record amp should not clip though.

 

[zonkers31]

You can also narrow down the problem by swapping cards with a good channel and a noisey channel. I believe there are 4 per channel. Sart with the repro card, and test. do this one card at a time and notice if the noise moves or stays on the channel. If it moves, you found your problem card.

 

[Pat Maki]

Calibration Tape?

Sometimes a calibration tape is not available or old and unreliable.

Have MRL tape, but am interested to learn about this method. Apologies for all the questions.... I do want to understand this.

You may check the gain of the repro amps with driving the recording into audible/visible saturation

So....increase output gain of mic preamp until monitored repro signal sounds saturated/VU is pinned....is that what you mean?

and check if these distorted repro levels are halfway equal on repro. 400..500 Hz seemed convenient to me.

Sorry, not following your meaning on this. Can you please elaborate?

The record amp should not clip though.

Are you referring to the record amp in the VU or the Record card? How does one measure what you're referring to? meaning...the VU only lets us see Input/Sync/Repro levels.