UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Vintage Audio (record players, hi-fi etc) (https://www.vintage-radio.net/forum/forumdisplay.php?f=14)
-   -   Technics SU-V60. Voltage regulator (https://www.vintage-radio.net/forum/showthread.php?t=155515)

samjmann 6th Apr 2019 6:59 pm

Technics SU-V60. Voltage regulator
 
1 Attachment(s)
Hi all,

I've been looking a slightly knocked about example of this amp. To protect the now hard-to-replace SVI4004 current amplifier, I've removed it from the amp whilst testing. On first examination pins 1&2 of the SVI4004 had been shorted together on a previous repair. One transistor in the voltage amp is short circuit (Q417). What is really puzzling me is the operation of the LT regulator, I've attached a snippet of the circuit below. The gate/source of each FET are shorted. What's the function of these FET's. Is it some type of electronic smoothing to cancel line ripple?

I've got 0v on the emitter of each regulator with a 48v+/- input. All the transistors check out OK. No heavy loads on either 18v +/- rails.

I've read on some of the audio forums that this amp is of the 'current-dumping' type, of Quad inspiration no doubt. It's supposed to be a pretty good amp, the figures look quite good for it.

The full circuit is on Hi-Fi engine, the snippet is just the regulator

Thanks for any advice on this.

SJM.

Radio Wrangler 6th Apr 2019 7:47 pm

Re: Technics SU-V60. Voltage regulator
 
The FETs with the gate-source connected are constant current sources. They sometimes get drawn as current source symbols or as diodes. They used to be made in a range of current values.

Theoretically you could do this with ordinary FETs but the variation in Idss from part to part is a major problem. These parts are from a selection process.

David

samjmann 6th Apr 2019 8:40 pm

Re: Technics SU-V60. Voltage regulator
 
Thanks RW for the reply. As I have no base voltage on either Q701/2, I suppose that the FET's must O/C. The FET'S are both 2SK246 not expensive! I'll replace both the 33uf's as well. Both Q701/2 test perfect out of circuit.

Cheers, SJM.

orbanp1 7th Apr 2019 12:07 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

As David suggested the FETs act as current sources and they have quite a variation in their Idss parameters (It is the Idss value of the FET that gives the current of the current source).
The Zeners are 18V 370mW devices, so you would want not more than 5mA to 8mA across them. The base current through Q702 is negligible.
You should measure the Idss current of the replacement FETs before you install them. Just connect the gate to the source, feed the drain source from a bench supply with something like 5V - 10V, and measure the drain current.

Regards, Peter

samjmann 7th Apr 2019 4:07 pm

Re: Technics SU-V60. Voltage regulator
 
Thanks Peter for the help. The 2sk246 has a rated Idss of 1.2mA, this is what Cricklewood state on their site. On the circuit the original 2SK301 has a Idss of 20mA, again figures from Cricklwood. If this was fitted then it would put more than the 5-8mA advised into the 18v Zener. Perhaps Technics revised what they were fitting during production.

When testing should I fit a 470ohm in series with the bench supply when testing to give a little protection to the FET?

Thanks again for the advice. :)

SJM.

orbanp1 7th Apr 2019 10:21 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

As mentioned, the FETs have quite a variation in their parameters.
Generally FETs of the same model number are ranked according to their Idss values. Do check the datasheet.
For the 2SK301 the Idss groups are:
P: 1ma - 3mA, Q: 2mA - 6mA; R: 5mA - 12mA; S: 10mA - 20mA.
For the 2SK246:
Y: 1.2mA - 3mA; GR: 2.6mA - 6.5mA; BL: 6mA - 14mA.
So pick the FET accordingly.

The 470 Ohm resistor would just drop some voltage, the FET should not be damaged without it when you measure the Idss.

Regards, Peter

samjmann 8th Apr 2019 10:10 pm

Re: Technics SU-V60. Voltage regulator
 
Thank you Peter for the information. I can see now that there is quite a wide variation in spec for the same transistor, so careful choice is needed!

I'll post again when I fit the replacements. These are 1.2mA-3mA Y class 2SK246's.

Regards, SJM.

samjmann 14th Apr 2019 6:54 pm

Re: Technics SU-V60. Voltage regulator
 
1 Attachment(s)
I've fitted the two FET's and both the supplies have come to normal levels. In the Voltage Control amp, Q417 was s/c collector-emitter, it's 2.2ohm emitter resistor had risen to 6ohms. These have also been replaced. At the mid point of each Voltage control amp I have now 6.2V. This should be 0v, but as both channels exhibit the same characteristic with the power amp Ic removed, I can only assume that this is correct.

Next step, to refit the power amp IC, then gently power on with a lamp limiter and variac...

SJM.

orbanp1 15th Apr 2019 12:18 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

I could not find a datasheet for the SVI4004, but here is a reverse engineered circuit of the chip:
https://elektrotanya.com/svi4004_sch.pdf/download.html

The circuit (together with its driver in the SU-V60) looks funny at the first glance, but I have not really spent any time on it to understand its working and to see if those 6.2V voltages in your circuit are correct or not.
Best course of action is to plug it into a simulator and see how does it work. (I might just do that.)
No chance to burn down anything in there!

Regards, Peter

samjmann 24th Apr 2019 3:45 pm

Re: Technics SU-V60. Voltage regulator
 
After a break because of work and the holidays, I've finally had a chance to look at this again. After a very careful power-up with a variac and a lamp limiter - both channels came up and the speaker relay engaged. ;D There is of course a 'but'... The defective channel (left) has a -68mV offset at the midpoint of the voltage amp, this is also present at the main speaker output midpoint too. The good channel is at -3mV, whilst not perfect is probably OK.
When I was first working on the amp just to get the unit somewhere near working as Q417 was short, I simply replaced it with an off the shelf replacement. I'm thinking that perhaps I should replace both Q415/417 with a matched pair.

I know the answer is to swop the transistor with it's R channel counterpart and if the fault moves, then the transistor must be the problem. The problem is working on the unit without causing any further damage. Both channels 'sound' OK at this stage, so at least the output stages have survived! The amp is so fragile when apart, probably the safest thing is to cough up and replace the pair in one go, without taking the unit apart unnecessarily.

Thanks for all the help so far. :) SJM

orbanp1 24th Apr 2019 11:25 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

I do not think that your problem is an "unmatched" Q415/417!

I mentioned the amp schematics looked funny on that reverse engineered circuit, well, it is just part of an amplifier.
The complete power amplifier also includes IC401, the "voltage control amp" and "power amp/current drive amp" are just the power out stages.
The amplification and negative feedback is done at IC401, which is a fancy op-amp, it is high voltage, low noise, dual op-amp, with emitter follower outputs.
IC401 is an AN7062.
Get the Japanese language datasheet, it has an application circuit example in it, the English language datasheet does not include that.
With that op-amp, the amplifier has a "Lin-type" architecture, just like a power op-amp, with all those buffer output stages.

If there is unbalance in that amplifier, and as it is a feedback amplifier, it could come from any stage in that circuit.
The DC output voltage of such an amp (think of it as a power op-amp) is set at the positive input of the amp, in this case the "+" input of IC401. It is set to ground through R403/404.
The AC/DC feedback to the "-" input is from the output of the amp.
Do measure the voltage at the "+" input point, the schematics shows -0.2V. It is probably from the input bias current of the op-amp. The "-" input of the op-amp also shows -0.2V.
Do measure and compare both sides.
If the voltage is different, it could come from a leaky C401/402 capacitor.
Disconnect that capacitor and measure the output DC offset again, does it change?
The op-amp itself could be faulty too. Unfortunately it is not a generic audio op-amp, or at least I do not know of any other such IC.
And it could be any other component in the amp.
Do measure all the transistors too.
And of course it could be a faulty final power IC as well.

The strategy to debug such a feedback amplifier is to check every active component and/or replacement of the parts that can not be verified by simple measurements (the ICs).

68mV offset itself is not that much, you can live it, but such amps usually have less than 10mV.

Regards, Peter

'LIVEWIRE?' 25th Apr 2019 9:01 am

Re: Technics SU-V60. Voltage regulator
 
Recently I've repaired a Cambridge A3i Amp. in which the DC offset in one channel could not be set below about 55mV, whilst the other channel's offset was below 4mV. Despite this, the A3i worked very well on both channels whilst on extended test and is now back with it's owner. No doubt the higher offset could have been reduced by changing pre-driver transistors, or other parts (The A3i uses a totally discrete transistor circuit, b.t.w.), but, as it continued to work well after the original fault had been fixed, I decided to leave well alone.

samjmann 25th Apr 2019 9:02 pm

Re: Technics SU-V60. Voltage regulator
 
Thanks Peter for such a detailed post. The AN7062 is on the front panel and accessible. I'll carefully measure both sides to check if there is any difference L-R.
On the current set up test point, the centre pin (L) is at -60mV. Whilst the R side is only 3mV. The power IC follows these voltages and the speaker midpoints are the same (virtually) as the test point. So I was thinking that the power Ic was probably OK.

I'm not using a meter probe round the power IC, I've had too many slipped probes in the past and subsequent bangs!

I'll get the unit out again soon and report back. Thanks also to LIVEWIRE? for your post. I too have had amps new, and not reported faulty with +/- 70mV on each channel output. I had to conclude that this must have been deemed OK by the makers. If all else fails I might have to leave it this way, but 'engineers hat on' we know it's not right.

Cheers. SJM :thumbsup:

samjmann 27th Apr 2019 7:34 pm

Re: Technics SU-V60. Voltage regulator
 
Just to update:

Did some careful checks on the AN7062 IC in the op amp or long tailed pair type circuit: Pins 2,3 (R good channel) -177.8mV
Pins 16, 17 (L defective channel) -238mV

The left speaker output has an -68mV offset. The right is about -3mV

If you subtract 68 from 238 you are left with 170. Not far off the correct voltage ref of -177mV.... Hmm..

I measured the two ref inputs to ground: Right is 100.5Kohm. Pin 2
Left is 103.7Kohm Pin 17

I can only assume that R403 must have gone slightly high?

I'll get a pair of 100k Hi stabs on Monday, switch on and hope!!!

Cheers... SJM.

orbanp1 28th Apr 2019 1:48 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

I do not think that R403 at 103.7k is your problem, that value is well within the ballpark!

The current through that resistor is higher than in the other channel, that is why it has higher voltage on it.
The higher current could come from higher bias current from the IC input, that means an out of spec IC. That would be bad news, unless you have a replacement IC.
The higher current could also come from a leaky C401 electrolytic capacitor.
Have you disconnected that cap and made a measurement? If not, do that!

Regards, Peter

samjmann 28th Apr 2019 7:00 pm

Re: Technics SU-V60. Voltage regulator
 
Hi Peter, :thumbsup:

Thanks again for continued guidance. As you say, 3.7K drift on a 100K is only 3.7% - within 5% spec. On the assumption that the source resistance of the IC's pin doesn't change with 100K to 103.7K then the current through the resistor would drop by about 7uA. I've got 238mV across 103.7k so that's a current of 2.2uA through the resistor

Using the good channel as guide: 0.177mV/100K = 1.77uA.
The faulty channel therefore is putting out 0.43uA more than the good.

I can't quite see why a leaky C401 would yield a higher resistance to ground,
but certainly if there was a little bit of DC leaking through that would certainly upset things. I'll check that tomorrow also.

The AN7062 is available from America at around £15.00 incl shipping, so if tight budget allows it could be replaced.

Regards, SJM.

orbanp1 28th Apr 2019 8:46 pm

Re: Technics SU-V60. Voltage regulator
 
Hi SJM,

It is not that the resistance increases from a leaky capacitor, but it can give some extra current through that resistor, depending what is the voltage at the other end of that capacitor. Of course, this assumes that the IC has the same input current on both the plus and minus inputs.

You could check a couple things.
The other side of that capacitor C401, the DC voltage should be at zero, or close to it. Do measure it, compare with the correct side (with no input audio signal).
Looking at the schematics, it is not likely that the DC voltage would be different from zero, but check it, none the less.
Do check if the the cap is installed with the correct polarity. Again, if the DC voltage across that cap is zero it should not matter.

If you remove C401 that will tell you if the extra current through R401 comes from the IC or through that cap. That would verify if the IC is fine or not.
You could definitely measure 7uA difference in input current, a DMM usually has a 200uA range.
Remove the IC, install an IC socket (two strips of those machined type IC connectors), and you can seat the IC in the socket with the pin in question bent up a bit, not inserted in the connector. Then you can measure the current.

Regards, Peter

samjmann 29th Apr 2019 9:50 pm

Re: Technics SU-V60. Voltage regulator
 
Hi all,

Fitted the replacement 100K, in fact selected one to give 99.7K. Switched on and the current drive did not light. The right channel then went to -57V 8-o
This turned out to be a dry jointed Q402. Resoldered, and current drive restored.

The output midpoint has now dropped from around -68mV to -56mV. Still not right, but a little better. Warming the AN7062 with a hairdryer (and nothing else) the midpoint falls to -49mV. It's looking like the IC may be faulty as Peter has suggested. As the input would appear to be high impedance, the slight change of 4Kohm does seem to affect the output of the IC. I haven't got a IC socket, but from what I've checked it would appear that the left channel is putting out slightly more than the right on the non inverting inputs. I did remove the 3.3uF capacitor, this didn't change anything and tested OK.

At this point I'm going to leave the amp playing for an extended period and see what happens. If the output stays at around -56mV or less I'll consider it fit, if not perfect. Thanks again Peter for all the advice, I've learnt a lot!

SJM


All times are GMT +1. The time now is 10:57 am.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.