NAB Tasscam 38

[evm1024]

I had thought that all tascam 38 (and all 1/2" 8 track for that matter) were IEC1. With that assumption I have been trying to calibrate my 38 to IEC1 to no avail. Rechecked a few times and I still got a 3 dB too hot playback at 16 kHz.

I went ahead and calibrated it up to NAB and indeed it played back +-1 dB from 60 Hz to 20 kHz (relapped heads are nice)

Digging around in the schematic and on the channel cards I see that C127 is listed on the BOM and schematic as 0.0039 uF but is stuffed 0.0022 on the channel card.

Anybody have a channel card from a 38 that they can check the value of?

Anybody hear of a NAB 38?

Regards, Ethan

PS C127 sets the record HF EQ.

 

[RRuskin]

I had thought that all tascam 38 (and all 1/2" 8 track for that matter) were IEC1. With that assumption I have been trying to calibrate my 38 to IEC1 to no avail. Rechecked a few times and I still got a 3 dB too hot playback at 16 kHz.

I went ahead and calibrated it up to NAB and indeed it played back +-1 dB from 60 Hz to 20 kHz (relapped heads are nice)

Are you using the machine's meters to calibrate? If so, be advised that they are frequency weighted and not accurate for calibration. You need an outboard unweighted meter unit to tweak things properly.

 

[Beck]

I've never heard of a NAB 38, but the channel cards are interchangeable with the 32 and 34, which are NAB.

 

[evm1024]

Not stuffed

These cards are 38 cards as that they are not stuffed for 7.5 ips.

32,34 and 38 cards are the same but remember that they are not interchangeable with out checking the eq values. (now where did the posting on the 3 cards EQ I made go?)

This is all very odd as that it does cal closely to my NAB MRL tape but perhaps not close enough to be called a NAB deck.

The 3 to 4 dB boost on playback of self recorded 16 kHz tones is what is throwing me. It could be how I m using the conversion table. More testing required and perhaps a few more email to Jay McKnight.

RRuskin, I'm not using the decks meters for calibration. However they read flat from 250 to 16 kHz and drop 1 dB at 20 kHz. I am using a not too bad DMM that has been tested against my Tek 465 for frequency response. Not 1% accuracy but then again not 30% error.

--Ethan

 

[Beck]

+/- 3 dB (40 - 22 kHz) would be within spec for that machine. But Tascam listed specs are generally conservative so I'd expect to see maybe +/- 2dB. That's what I have with my TSR-8 anyway at 40 - 20 kHz.

Yeah, the interchangeability thing depends on how the cards are populated. I mentioned it in the off chance a previous owner used cards from a 34 or somethin'.

 

[RRuskin]

These cards are 38 cards as that they are not stuffed for 7.5 ips.

32,34 and 38 cards are the same but remember that they are not interchangeable with out checking the eq values. (now where did the posting on the 3 cards EQ I made go?)

This is all very odd as that it does cal closely to my NAB MRL tape but perhaps not close enough to be called a NAB deck.

The 3 to 4 dB boost on playback of self recorded 16 kHz tones is what is throwing me. It could be how I m using the conversion table. More testing required and perhaps a few more email to Jay McKnight.

RRuskin, I'm not using the decks meters for calibration. However they read flat from 250 to 16 kHz and drop 1 dB at 20 kHz. I am using a not too bad DMM that has been tested against my Tek 465 for frequency response. Not 1% accuracy but then again not 30% error.

--Ethan

How much over-bias are you using and @ what frequency are you using to set it?

 

[evm1024]

per spec and ear

How much over-bias are you using and @ what frequency are you using to set it?

I'm using 5 dB overbias set by ear with a 10 Hz signal for lowest distortion. Testing at 16 kHz gives 5 dB over.

One can get the high freq response down to spec by using 10 dB over bias. This of course gives loads of distortion and would most likely overload the bias amp should all 8 channels be set the same.

I should be able to get some time today to work on the deck again. I'll be shooting for a +-2 dB response rather than a flat response....

 

[RRuskin]

I'm using 5 dB overbias set by ear with a 10 Hz signal for lowest distortion. Testing at 16 kHz gives 5 dB over.

One can get the high freq response down to spec by using 10 dB over bias. This of course gives loads of distortion and would most likely overload the bias amp should all 8 channels be set the same.

I should be able to get some time today to work on the deck again. I'll be shooting for a +-2 dB response rather than a flat response....

1. Before setting the bias, have you put the hi eq pot at mid point?
2. How can you determine 5 db by "ear?"
3. My recollection of Tascam decks was that they were happiest with a 3 db overbias @ 10khz. I cannot remember if they wanted the input level to be -10 or 0 VU.

 

[evm1024]

1. Before setting the bias, have you put the hi eq pot at mid point?
2. How can you determine 5 db by "ear?"
3. My recollection of Tascam decks was that they were happiest with a 3 db overbias @ 10khz. I cannot remember if they wanted the input level to be -10 or 0 VU.

1) did not set the pot to mid point. Left it at the last cal point.

2) Aren't your ears calibrated? Mine are not also.. But I can go back and measure the voltage differences to find the overbias so that I can set it by my DMM.

3) The spec calls for 4 to 5 dB overbias at 10 kHz at -10 VU. (I follow spec unless I know otherwise, so I'm sure that I set the bias at 10 kHz and not 16).

This 38 is new to me but not my first deck.....

 

[evm1024]

Success!

Time to report success!

What caused success is less obvious....

I did pull all the channel cards and put contact cleaner on all the sockets and pins, Also put pot cleaner on the pots. Next I pulled my sticky MRL out and baked it for a while (20 hours....).

Then I calibrated playback very carefully and did not change meter scales. I should note that in prior calibrations I set the repro cat 1 kHz at 0 VU (0.316 mV) and EQ at -10 VU. This time I set repro cal at -10 VU (100 mV) and repro EQ also at -10 VU.

Which of these differences was the problem? Hard to say.

End result is +-2 db from 30 Hz to 20 kHz.

--Ethan