TSR8 DBX faulty channel, but signal present if DBX switch bypassed

[carausius]

I've got this long standing issue with my Tasacam TSR 8 channel 4 card, the signal plays back fine with DBX switched out but when DBX is switched on there is no output signal present. The other 3 DBX channels work fine on that DBX switch.

I took the card out & swapped it over with another channel but the exact problem followed the card with the replacement card working fine on the same channel. I cleaned the card with PCB cleaner / isopropyl alcohol with no luck.

I looked at the schematic in the DBX section of the r/p card circuit below.



I assumed the problem lay inside the U10 switching circuit, so I removed & socketed the U3 & U4 op amps (NEC 4570C ) with equivalent LM833N but no change,. I removed & socketed U6 DBX VCA & also tested with equivalent THAT 2180 without any change.

There's 5 transistors in the DBX circuit serction, Q17-Q22 (2SD1915 ), I've tested them all & some appear to be faulty, although I tested in circuit so perhaps not?

Or the U10 (LC4966) 4way bipolar switch itself could be faulty?

Anyone with any advice on what to try next would be very welcome!

 

[witzendoz]

The board could have a faulty capacitor, even if they get out of spec they can cause problems.

Alan.

 

[SkywaveTDR]

I would not be blindly changing chips as a fix and I would not be using so called compatible IC's for the original ones in the unit. Why don't you place a 1KHz signal in and trace the signal as to where it stops? Capacitors are a number one item of failure on any card and in the case of SMD parts they are almost tripple the failure rate as SMD parts are always junk. Just like he 122 Mk II and Mk III capstan motors with a 100% failure rate. Put through hole caps on the card and then deck never has any problems again.

 

[carausius]

I can send a 1khz signal & supply say 15 + - DC to the card when it's removed from the machine, but I'm not sure how I would switch the DBX on & off. Maybe I could inject a signal after the U10 switching occurs in the circuit so bypassing that stage?

I'll check the schematic & the datasheets & see what I can find. I've just found my very basic sine /square wave generator https://www.lighthouseaudiovisual.co.uk/dap-audio-sc-17-sinesquare-70827-p.asp

So that's now working with my ancient analogue Hameg oscilliscope. Got a 1khz sine signal all through the DBX circuit section & to the output of the channel card. The channel works in playback without the DBX switched on, but not with DBX engaged.



Cheers John K

 

[SkywaveTDR]

Some of the noise reduction circuit seem to have signals disappear in amplitude and I have fixed many dBx units that have IC's. These are a lot easier than the Dolby S on the 238S where there is not even documentation for them.
Sometimes in service of these an extender cable or card needs to be used. There needs to be a control line that turns the dBx routing on. So far on all the TSR8 I have worked on they did not have dBx problems.

 

[carausius]

The service manual mentions an extension card for the DBX, seems that perhaps it's a switching issue.

I've been checking the connections to / from the LC4966 quad switch that does some or all of the DBX switching, voltage including via U11-U21 network which are digital transistors
http://www.cbtricks.com/radios/galaxy/datasheets/transistors/pdf/dta124ee.pdf

I can't even work out which header sends the signal voltage to activate the DBX switch!

U8 is a microcontroller, U9 CMOS NAND gate, lol, I think I'll have to pick up a replacement card! Can't find much information online on DBX circuitry or repair either.

 

[SkywaveTDR]

From the schematic I just checked if playback is the issue you need to look at the input and outputs and control circuits concerning U10. Determine if the input control signals are there then put a IC socket in the board and try a different chip IF you find something wrong with it.
It can be as simple as a bad digital transistor. It only take a DVM to test digital input signals to the U10 inputs.
The head signal goes in a pair of transistors Q1,Q2 as well as Op amp U1 which is a 4570 IC and comes out the Rep Pot to go to pin 4 of U10 which when on send it into the dBx circuit through SW1-2, if the SW1-2 is not closed in play mode which is what U10B make, the signal bypasses the dBx process in taking the signal through U10C as U10D will be turned off. The same dBx process is used for play and record and the heads are substituted for the inputs through these electronics switches. So if it records with dBx then that circuit is working but you have a lack of routing of the play signal all the way through. This is why a large print out of the schematic is beneficial as looking at two PDF file pages will drive you crazy as well as eye strain.

 

[carausius]

Finally got some 4570's so tried those on U1, U3 & U4 with no luck. Tried a new digital switch U10 LC4966 it's the same problem following the card, DBX doesn't switch on, another card works fine when substituted on the channel.

I checked the 2nd DBX chip U7 which I thought was a DBX 1252 VCA like U6 is, but it's actually a DBX 1253 RMS detector. One solder joint looks suspicious from one angle. Could be a dry joint but I'm sure I looked it over before?

I was tempted to just redo the joint but decided to socket it in case it needed replacing. Took a while to get it out, no leverage due to close components.

I fitted a socket, back in it went...........! We finally have DBX signal!!!

Thanks for the help guys, special mention for skywaveTDR & his valuable expertise!

 

[SkywaveTDR]

Wow all that for a bad solder joint? I have always said that these things are not soldered all that well requiring me to go over many joints in decks. At least you figured it out.

 

[carausius]

If you actually have to re solder any of the board, do you just reflow the joints, or remove the old solder beforehand & use new solder for the joints?

I assume these PCB's were dipped in solder trays from the underside? Can't see them doing it all by hand! There's a a bit of flexibility in them compared to modern PCB's.

 

[SkywaveTDR]

I have never seen them dipping boards in by hand. Over at Searl Analytic they had a conveyor that ran many boards over bubbling solder vats that when the solder made contact to the board, parts and flux it just stuck on where it was suppose to. The problem with this fast and cheap method is that too little solder makes a joint. It is nothing like those the solder training books tell you about. They would all be classified as wrong solder joints. But since it is consumer and made for profit they get away with it. Vibration and dropping or twisting the frame of a device can easily cause the joint to fail. The joints that go through many thermal cycles will break even before that. I have seen joints that are crystallized and therefore not given enough time to cool.

Yes, I add Kester 44 solder to as many as 200 joints in a deck and when I come across a joint that does not meet my specs then I suck off the old solder and it gets 100% Kester 44 solder.
There is a soldering job I call reliability soldering. This kind of joint will not break or crack as I believe it to be to Aerospace grade soldering as I was trained at Hughes and VCA Teletronics by a NASA Solder instructor. It all comes down to the right amount of solder applied and to make sure the solder bonds to the component lead. Some older components get a layer of black on them that black solder from sticking. This must be recognized and a wire brush used to clean off the leads or a aggressive solder paste.
Solder I do not like to use is lead-less and that from China that reminds me of slag.

 

[wkrbee]

Visual inspection is the first step in repairing almost anything.

 

[Joseph Macry]

I had a similar problem with my DX8 noise reduction unit. Turned out to be a shorted electrolytic cap (tantalum) between +15 rail and ground. I don't see such in your pic of the circuit diagram, but it is probably there, off to the left somewhere.