AKAI GXC-46D Service Manual

[ButchJames]

Hello everyone.

Somewhere within the forum, I've seen someone offering a copy of the AKAI GXC-46D service manual, not the schematic, which is easy to access.

Could I be a cheeky little so-and-so and ask if someone would be willing too offer me a copy?

By all means PM me if you're happy with that.


Thanks.
Eric.

[arjoll]

That'd be me. Drop me a PM with your email address and I'll send the PDFs in the morning.

[ButchJames]

Thank you!

[ButchJames]

Well, I received the AKAI GXC-46D this evening, and straight away I opened it up. After about 2 hours of fiddling about with it, the prognosis is -

Initial issues: (1) mildly dusty, (2) some corrosion, (3) sloppy main belt, (4) all play/ff/rw stops after 3-5 seconds, (5) higher noise on right channel on playback, (6) pinch roller needs replacing, (7) All support sponge material has disintegrated.

• Dealt with (1), (3), and other cleaning and switch cleaning.
• Can deal with (4) as I see the belt that triggers reed switch movement is slipping.
• Probably new transistors to deal with (5)?
• Pinch roller (6) looks straight forward to replace, I think I have the right one in stock!

Induction motor is in good nick, takes about 5 or more seconds to slow down! (linked with flywheel)

It will be interesting to see what thickness of belt I may need for the most precise 4.75cm/s speed!?
So far, I put on a belt which is about 105mm diameter and may be 1.5mmx1.5mm in CSA? This is all temporary, just to get the machine basically working.

So far, it's looking promising.

[ButchJames]

In my original post I forgot to mention that i had jut bought a GXC-46D off ebay - sorry for any confusion.

Well, so far I've changed the take-up spool belt. this was a little tricky, but it always is first time round. This did *not* cure the auto-stop issue, where it triggers automatically after about 3-4 seconds. So to stop this perpetually happening, I've disconnected the solenoid plunger and put a 300ohm (5W) resistance across to act as a load. I didn't have to do that - I just decided to let the final transistor cct think it's operating.

I did notice that on Switch-ON, my DAB radio breaks up its radio reception - that's quite easily done if there is a burst of EM radiation nearby. And I immediately got the impression that the noise/spark suppressor snubber cct was faulty? I don't have any direct replacements, but (for now) I've soldered in parallel 2x10nF 2KV rated caps across the switch. No EM bursts now!

The original spec: 33nF/120ohm snubber.

[ButchJames]

My final post before I continue to fix this AKAI.

For those interested -

The default pinch roller runs on a 2.5mm pin, and secured with an e-clip! Something I've not seen before? So a new roller with a 2.5mm centre will be on order.

Currently, I have the deck running on a 2mm rivet 'pin', with a 13mm x 8mm x 2mm pinch roller. Everything seems fine, but I will swap back to a proper 2.5mm pinned roller later when i get some.

As a quick check I ran several tapes without a pad to see if the path taken by the tape was stable - seems fine.

Eric.

[ButchJames]

The previous post was supposed to be my final, but it is worth mentioning that the problem is the noisy pre-amp is solved: replaced all head-amp 2SC458 with KSC1845-FTA (NPN), and the 2SC711 with KSC1815-GR. All working good, no noisy amplifier!

[Whaam68]

Good work I have a couple of these in the loft and they are really nice sounding decks as a lot of the early 70'd decks are but are generally overlooked for their lack of more "modern" features. One of mine had lost most of the red paint off it's VU meters and there's a good youtube vid from a guy in Switzerland showing disassembly of the meter movement which I followed and it came up really well.
Mike

[paulsherwin]

The 2SC458 is a Hitachi general purpose transistor widely used in 70s and 80s Japanese audio gear. There are lots of reports of them going noisy, but they are normally easy to replace with other GP types.

[ButchJames]

I recently bought two of these - the one above is almost back to full spec. I will later re-transistorise the Dolby circuit which should not be overlooked, as they are in operation even if you don't have Dolby switched-in. At least that is the case with these tanks.

To my complete surprise, I really am charmed by this machine. I recall it well back in the 70s, but as it was 'out of fashion' I didn't fancy it, and Dad and I bought the later GXC-75D in February 1977. (Funny how we recall some dates)

The latter 46D I received yesterday is also now working, it was full of dust internally, almost comical! That will be restored fully at a later date.

Pinch Rollers: I really need to get some 2.5mm-holed pinch rollers, I know where I can source some decent ones. For now, I have improvised and use a 2mm-holed roller and have made a small brass collar to help the pin remain stable and upright - I'm quite chuffed that it works so well. May even leave as it is!

The amount of decks I have is a little embarrassing, I think I'll have to sell a few off!

[ButchJames]

Originally Posted by paulsherwin

The 2SC458 is a Hitachi general purpose transistor widely used in 70s and 80s Japanese audio gear. There are lots of reports of them going noisy, but they are normally easy to replace with other GP types.

Thanks for the feedback.

There seems to be some '2SC458LG' s in the Dolby cct. Looking at the specs and how they are used, again I think I can resort to the KSC1845 and KSC1815 series, very probably the latter. The 1845 only requires very low quiescent currents, and is generally used for low signal amplification, so I'll chance the KSC1815-GR, they've always been good.

Current gain hfe at over 100 (certainly above 200) becomes almost immaterial as negative feedback mathematically dominates/controls the gain, not the hfe parameter. I've seen so many folk tie themselves in knots over hfe figures, they are blissfully unaware that this doesn't matter beyond a certain point.

[paulsherwin]

The 2SC1815 is another very common GP type in audio gear. They are roughly equivalent to the 458.

[ButchJames]

Just been studying the Dolby circuit this afternoon in a coffee shop, and examining the approximate bias/quiescent currents.

For now I'll make the following swaps:

Old 2SC458 > for the new KSC1815-GR
Old 2SC458LG > for the new KSC1845-FTA

There is a PNP (looks like a constant-current 'loading') transistor in there - I'll use the compliment to the C1815, that is, the KSA1015 .... I think I have some in stock!?

There is also an N-type JFET too to look at.

I like using the 1845/1815 series since their pinouts are often the same or the reverse (with collector in the middle) to the older equivalents.

A possible alternative is the BC547-BC550 series, but they have their pins configured differently, although they are good transistors. Very good Noise Figure specifications. I have used these many times.

Actually, the BC550 may have been a better choice at the front end of the head pre-amplifier. Anyway, for now I'll leave well alone!

[ButchJames]

The Dolby circuit has been re-transistorised as follows:

All 2SC1312 (not the 2SC458LG as expected) NPN transistors were replaced by KSC1845-FTA
The two 2SA658 (PNP) were replaced by the KSC1015-GR.
The other 2SC458s were swapped for KSC1815-GR.

All went well, exept that online data for the 2SA658 was confusing, and had me wiring them the wrong way at first.

A big downside to some early decks like this one, is that there appears to be no on-board component labelling, so a one-for-one, or two-for-two (L-ch R-ch pairing) approach is best? Anyway, it's all working now.

Regarding the solder side, I am mindful that any high impedance or voltage sensitive amp stage may be vulnerable to stray micro-discharges which can result in small, but perceptible transistor noise-bursts. So, to eradicate this possible scenario, I'll often apply a thin layer of clear nail varnish to the vulnerable area after soldering and cleaning has been done. Works a treat.

Back Tension:

This deck appears to have zero back tension provision? I was drawn to this when sometimes I notice a shift in azimuth when listening with headphones. Tensioning the tape by hand and azimuth is easily stabilised. I need to devise something that creates a small amount of drag for the back spool.

[ButchJames]

Back Tension: Just a note on the back tension point I made earlier. I decided to try out a few felt pads and this arrangement seems to work the best so far.

Four cut sticky-back felt pads one on top of each other which just stroke the base of the supply spool. No azimuth drifting now at either end of the tape. As a quick check without the oscilloscope, I recorded white noise at about -20dB (ref OVU) and listened in for a few minutes.

RW/FF not affected.

[ButchJames]

I may return to the back tension issue later, but for now it's good. The problem being - so little space to work on, and with that, the scope for ideas become a little constricted.

Record Bias:
Casually looking at the recording bias HF frequency amplitude at the head, and I can report that it is around 12v-18v peak. To be able to record/pb with a flatter frequency response and on Type I 'modern tapes', that bias is going to be somewhere from 18v - 22v peak IMO.

Bias Access:
The oscillator and bias board are as one. High frequency bias amplitude can be altered via the two 300K ohm (I think?) potentiometer underneath. Yes, that's underneath?!

Record/PB Levels:
Can be found on one side of the deck with the wooden surrounds removed.

Pre-emphasis:
AKAI have employed an unusual pre-emphasis circuit, split apparently into two parts. (1) Boost HF (as traditionally done) using a LC 'resonance' circuit in the emitter leg. (2) An additional Voltage-Shunt feedback loop taken at the output with a series LC 'trap' to negate gain gradually to a minimum at approx 1/(2??(LC)) Hz. This later addition was to offset high frequency saturation and hence HF distortion; but this was in the days when Normal tapes were a little inferior to the modern types, their SOL was lower.

I may alter this circuit a little later to extend the frequency response. So far I can achieve slightly better than 30Hz - 14,000Hz ±3dB, or a very flat response but at about 3dB down at 12,500Hz.

As the machine stands now, wow/flutter is around 0.11 .. 0.13% WRMS. And that is without tweaking the setup with a slacker belt drive. Tape speed is about 1.0% ... 1.3% fast.

Hope GXC-46D owners/users find some of this interesting.

[ButchJames]

I forgot to mention - there is no Type II bias, which is traditionally about 50% higher. This machine is old, and so Chrome tapes back then seem to be biased at the same bias level as for Normal tapes?, except that pre-emphasis (via circuit '(2)') was a little different if we study the circuit. And from the same period, I have a Sony TC-134SD (and a sack full of TC-161SDs) where both decks only offer the same bias and PB equalisation ('de-emphasis') for both tape types.

It appears also, that de-emphasis is the same for both Normal and Chrome.

Any comments welcome, or even corrections to what I've written.
Cheers.