Tascam 234 improvements?

[Bo Deadly]

I have a Tascam 234. It's a cassette machine that runs at 3 3/4 ips. Supposedly it's the best of the cassette machines.

I have performed a preliminary test of each channel by recording a test signal and then looking at it in the QA400 analyzer. This is just line in to tape and then playback through line out. No DBX. Test signal was confirmed on the scope to be -10 dBV (which is 0 dB on meters) and playback was adjusted so that the QA400 read -10 dBV.

Here are the results:

How does this look for a tape machine? Is horrible or just mediocre? How would this compare to a high quality reel-to-reel? Am I 10 dB off or 20dB or what?

The manual claims playback SNR should be 52 dB without DBX, unweighted, 20-20k.

Although, I'm a little more concerned about audible transient issues. There are sudden changes in volume (DBX is out) and fluctuations in hiss. The 3rd harmonic is pretty serious.

I was going to just try to replace caps (starting with PS) and see what happens but I thought I would ask around before I burn through a bunch of electrolytics only to find that all of the issues are ultimately just limitations in trying to use a cassette / tape.

If this is worth working on, what would you do to improve performance?

Would checking oscillators, traps, demagnetizing, ... etc improve things appreciably?

Here is the schematic:

https://groupdiy.com/index.php?action=dlattach;topic=65910.0;attach=50618

 

[ruffrecords]

You have basically got 1% third harmonic distortion which is not unusual for tape at nominal operating level. Headroom and S/N for tape is usually specified for 3% distortion which would be 10dB higher than you have measured. What S/N does the quantum give?

Cheers

Ian

 

[benb]

I have a Tascam 234. It's a cassette machine that runs at 3 3/4 ips. Supposedly it's the best of the cassette machines.

I have performed a preliminary test of each channel by recording a test signal and then looking at it in the QA400 analyzer. This is just line in to tape and then playback through line out. No DBX. Test signal was confirmed on the scope to be -10 dBV (which is 0 dB on meters) and playback was adjusted so that the QA400 read -10 dBV.

Here are the results:

How does this look for a tape machine? Is horrible or just mediocre? How would this compare to a high quality reel-to-reel? Am I 10 dB off or 20dB or what?

The manual claims playback SNR should be 52 dB without DBX, unweighted, 20-20k.

Although, I'm a little more concerned about audible transient issues. There are sudden changes in volume (DBX is out) and fluctuations in hiss. The 3rd harmonic is pretty serious.

I was going to just try to replace caps (starting with PS) and see what happens but I thought I would ask around before I burn through a bunch of electrolytics only to find that all of the issues are ultimately just limitations in trying to use a cassette / tape.

Shotgunning the electrolytics might help (especially if you want this thing to work in future decades, but the mechanics may also be iffy about lasting that long), but most of this seems to be limitations of the cassette format.

If this is worth working on, what would you do to improve performance?

Keep the heads and tape path clean, try the 10 best cassettes you can find.

I recall the early-mid 80s, recording and playing a tone on an "ordinary" (forget the model, but not high-end) average stereo component cassette recorder on the "standard" TDK SA-90 and watching it play back on an oscilloscope. The wave would wiggle due to wow and flutter, and maybe every few seconds it would drop down to less than half amplitude for a couple cycles (dropouts). I got some Denon tapes that were better (the SA had been "the best" when it was introduced, probably a decade earlier).

Do they still make "high-end" blank cassettes? I read where the pre-recorded cassette market has come back, but don't know if that tape would be as good as the best available from decades ago.

If you want tape, IMHO the next step up would be reel-to-reel. I still have a 3340 (because I couldn't sell it 20 years ago) it's far from the best quality 4-track, but it's (mostly) better than cassette, and amazingly it's worth something again! The current problem seems to be affording tape.

 

[JohnRoberts]

The 234 may be as good as it gets for 4 track cassette recorders, Nakamichi is (was) regarded SOTA for 2 track consumer format cassette.

The 3 3/4 IPS is 2x as fast as the 1 7/8 IPS 4T cassette I did at Peavey.  2x speed means two times the magnetic domains running past the head so better S/N. 

I am not a fan of 4T cassette but perhaps because I was responsible for making one not suck...(too bad). Peavey's deck used dolby C which was a major PIA for set up...  That cassette had more trims and factory adjustments than 10 other Peavey products combined.  I had two technicians working full time doing final set up, and they were lucky to get a handful out a day.

The heads in those had guard bands between each track so you could perform punch-ins to one channel at a time... making the tracks a little narrower than consumer format.

Did I mention it was a PIA?

JR

PS: One hook the Peavey deck had was external speed control so it could be synchronized to a VCR using SMPTE time code, but burning one track of a 4T for time code made it even more marginal for serious use.. : That SMPTE machine synchronization project was a major PIA too...

 

[Bo Deadly]

You have basically got 1% third harmonic distortion which is not unusual for tape at nominal operating level. Headroom and S/N for tape is usually specified for 3% distortion which would be 10dB higher than you have measured. What S/N does the quantum give?

I have raised the level sufficient to make ~3% THD in the recorded output. Gain structure was 0 dBV out -> switch to record -> adj input level so that QA400 read ~-5dBV -> switch to playback -> assert ~3% -> adj output level to not clip QA400.

Here are the results:

The QA400 claims the SNR is ~41-44 dB (noted in legend). So it seems I'm about 10 dB off to spec'd SNR of 52 dB.

Just for comparison, the manual for the Tascam 22-4 reel-to-reel has an SNR of 55 dB unweighted no dbx 7.5 ips. Surprisingly not much better than the Tascam 234 and it's track width is 0.9mm vs 0.5mm respectively.

I wonder how much that 3rd harmonic is affecting the SNR.

I guess I could try changing the filter caps and then run the same test again.

 

[ruffrecords]

I wonder how much that 3rd harmonic is affecting the SNR.

I guess I could try changing the filter caps and then run the same test again.

The reading you have is probably most;y distortion. Now you know that 0dBu represents 3% harmonic distortion you can just measure the tape noise with no signal to get the signal to noise. That will be how it will have been measured by Tascam.

I still have a Fostex R8 that squeezes 8 tracks onto quarter inch tape which means the track widths are the same as a cassette. Its unweighted S/N is specified at 40dB and it runs ar 15 ips.. The Dolby C takes that into the 70dB plus range. I have made some lovely recordings on it. It has a great sound.

Cheers

Ian

 

[Bo Deadly]

Unfortunately it seems 0dBV is not 3%. If I use 0dBV I get:

R/P   Inp  THD   SNR
rec   0dBV 0.02% 88.7dB
ply   0dBV 10.2% 42.6dB
ply -10dBV 11.1% 41.8dB

The first row is 0dBV output on the QA400 and 234 input level adjusted to yield 0dBV on the QA400. This is recorded to tape but of course what the QA400 is seeing is the "pass through" signal.

The second row is the playback of that recording with the output level again adjusted to read 0dBV on the QA400. This shows that 0dBV gives much higher THD than the target 3%. I *think* 0dBV into the 234 is supposed to be 3% though.

The third row is the 234 output adjusted for -10dBV on the QA400. This shows that the output level does not affect the results much.

So after some trial and error I found that if I adjusted the 234 input level to -5dBV, that would get me vauely close to 3% when played back. Here is the plot and corresponding table using -5dBV:

R/P  Inp  THD   SNR
rec -5dBV 0.02% 83.8dB  
ply -5dBV 2.8%  44.3dB

The first row (blue trace in plot) is 0dBV output on the QA400 and 234 input level adjusted to yield -5dBV on the QA400. Again, this "pass through" signal is also recorded to tape.

The second row (orange trace in plot) is the playback of said signal with the output level adjusted to read -10dBV on the QA400. I used -10dBV because that is the nominal level for the 234 (at 0dBV the meter needle is violently thrashing).

The grey trace is the same recording but with the QA400 stimulus stopped. Meaning no signal.

Unfortunately it's not crystal clear to me how to go about determining the SNR in this case. It's not even clear that this is a valid experimental process since the QA400 might be using the signal being output to calculate SNR (which is not used during playback from tape). In fact, sometimes I have to stop/start the QA400 to yield a sensical spectrum. But when the level is stable it does seem to work.

One part of the circuit that catches my eye is the recovery amplifier which is U102 in the schematic:

U102 RC4558P
Q101 2SK270A GR

According to the block diagram in the manual it's making almost 50dB of gain. I wonder if replacing C109 would help get be closer to the spec'd 52dB SNR that clearly I'm not getting right now. Then maybe replacing the recovery amp altogether might be worthwhile.

 

[ruffrecords]

OK, I take it you are using the line input and line output. Feed your 1KHz into the line input and set the input control so the VU meter reads 0VU ( the user manual states the peak recording level is 8dB above this). Record some tone. Replay this recording and set the output level so it reads -8dBu on the QA ( to account for peak level being 8dB higher). Leaving the controls set the same now replay some blank tape into the QA. The QA should give you a level for the residual noise in dBu. This value will be the S/N ratio relative to peak recording level.

You cannot make the measurement with signal present because the distortion is greater than the noise.

Cheers

Ian

 

[Bo Deadly]

OK, I take it you are using the line input and line output. Feed your 1KHz into the line input and set the input control so the VU meter reads 0VU ( the user manual states the peak recording level is 8dB above this). Record some tone. Replay this recording and set the output level so it reads -8dBu on the QA ( to account for peak level being 8dB higher). Leaving the controls set the same now replay some blank tape into the QA. The QA should give you a level for the residual noise in dBu. This value will be the S/N ratio relative to peak recording level.

You cannot make the measurement with signal present because the distortion is greater than the noise.

0 VU on the 234 is -10dBV on the QA400 so playing back at -8dBV is +2 VU. I think -2dBV would be +8 VU.

But I'm not sure it matters. Unfortunately the QA400 does not display any noise values with no signal:

 

[ruffrecords]

But I'm not sure it matters. Unfortunately the QA400 does not display any noise values with no signal:

Yes it does. I know, I used to have one!!

Cheers

Ian

 

[radardoug]

Are you trying to obtain good audio quality from a cassete? Dont. With todays digital gear exceeding 100 dB sig/noise, it just shows up how bad the cassete formats were. If you want tape quality, buy a good tape machine, and put the cassete in the place it always should have been, the rubbish tin!
Sorry, thats the reality!

 

[enginefire]

An alternative is to hit the tape nice and hard (to your taste and style) then transfer it to the computer and null out the noise with software.  Not ideal but it may improve your overall results.

 

[Bo Deadly]

Are you trying to obtain good audio quality from a cassete?

No. You are correct. This is definitely a lo-fi instrument. It has a thick sound which I assume is a function of wow and flutter and also the cause of bad SNR.

Although one thing that I really like about the Tascam 234 is that it's stand-alone rack mount. If I'm having a creative moment and I want to record it quick, I don't want to wake up the computer and launch the capture program and create a new track and click the drop down with the little triangle and then scroll to the thingy ...

I had a Zoom R16 once and it worked beautifully but all of the "digital multitrack recorder"s are table-top with builtin mixer, eq knobs, effects and other stuff I don't need or have space for.

If there were a stand-alone SD card recorder but with controls like a reel-to-reel for easy overdubs I would definitely be interested in that. None of the current rack mount digital recorders can do overdubbing - you can't simultaneously play and record arbitrary tracks and "bounce" tracks and so on.

There should be a 1U rack mount recorder with 8 sets of input select/level/pan controls with balanced dsubs on the back for line in and line out. The process should be just like an old reel-to-reel. Hello Tascam?

The Alesis HD24XR would be closest to the sort of thing I'm thinking of.

 

[radardoug]

Fostex D90/ D160/ D2424. Just like a tape recorder, only digital.

 

[Bo Deadly]

Can someone advise me regarding bias adjustment?

The oscillator is putting out a nice sine wave with amplitude of 145Vpp. I measured the bias level after the adjustment screw (180p tunable cap) and at TP1 which is just after the record driver but before the 100kHz trap that isolates the OA from the rec/ply head. Here's the values:

ch1   bias level TP1
1      80Vpp     1.8Vpp
2     100Vpp     2.1Vpp
3      98Vpp     2.7Vpp
4      72Vpp     2.1Vpp

Does this seem reasonable? Should the bias levels really be so different?

Are these traps and bias levels something I want to mess with?

Can I adjust the bias level with just a scope and an audio analyzer (QA400)? I do not have any test tapes or anything like that. There's nothing in the manual about bias adjustment in the Tascam 234.

What if I just record and simultaneously adjust the bias level +5Vpp every 10 seconds and then play it back through the audio analyzer and see if the response / noise get's better / worse, etc?

 

[ruffrecords]

The 'normal' way to bias tape is 1dB over bias at 10KHz. What this means is, on a three head machine, you record a 10KHz signal, usually about 10dB below normal operating level. Then you sweep the bias through the available range from min to max and observe the level of the 10KHz playback signal. As the bias is increased, the signal will increase until it reaches a peak and begins to fall. You adjust for 1dB fall after the peak. This provides a reasonable compromise between frequency response and distortion.

if you don't have separate record and replay heads this is nigh on impossible.

Cheers

Ian

 

[pvision]

There are quite a few rackmount audio recorders out there. Joe Co make some high-end ones but there are others, maybe not rackmount, from Zoom, Tascam, Cymatic & etc

One of the RME interfaces (UFX?) has a USB slot which can be used to record multitrack audio directly off the preamps and mixer in the interface. That's about as minimal a setup as you could get

With a cassette transport you're always going to be limited by track width & speed variations and those aren't limitations you can work around. I can't help feeling it's time wasted trying to get decent results from it

Nick Froome

 

[Di_Sousa]

This is definitely a lo-fi instrument. It has a thick sound which I assume is a function of wow and flutter and also the cause of bad SNR.

Nice that you have that unit, it sounds really nice indeed.

I used to have one, and I still love the Drums sounds I recorded using it. I used to also use it in mixing, I would send tracks from Pro Tools to the Tascam when I needed that particular sound, and things sounded too sterile for my taste.

For me it's not a matter of Lo Fi or not, it's it sounds a certain way and I like it and it's usable to me.

Unfortunately my unit was stolen when I live in the UK. I also loved that it was in the Rack format.

Nowadays I have an Akai MG614, it's a really nice piece of gear, sounds good also but different than the Tascam 234.