DAC90A Front End Woes.

[David G4EBT]

The DAC90A is such a commonplace radio that it’s hard not to trip over one at any car boot sale, hence, it’s almost become a ‘rite of passage’ for many newcomers in the hobby, despite it not being the ideal set for inexperienced restorers due to it being a live chassis set.

Like many forum members, I’ve lost count of how many DAC90As have passed through my hands over the years that I’ve successfully restored. Hence, I’m more than a little embarrassed to say I have one in which the frequency changer fails to operate and has me stumped. I started restoring the set about five years ago. It was a car boot acquisition and a non-worker, though hadn’t been tampered with. Having replaced several out of spec resistors and waxy paper caps, and dodgy wiring the AF/RF stages worked fine.

I set the signal generator to 465kHz and in line with the Bush datasheet, injector a signal into V2 control grid (pin (6) to adjust IFT2 if need be. Didn’t need to – it was spot on. Likewise 465 into V1 control grid (pin 6) showed that IFT2 needed no adjustment.

However, I couldn’t get the frequency changer to behave, so when other priorities took over, I shelved the set intending to return to it sometime, whenever. Fast forward five years and I’ve got it back on the bench, thinking ‘how hard can it be – it’s just a DAC90A?' Well harder than I imagined! The annotated diagram of the front end highlighted in blue shows all the connections that I’ve checked for continuity. The LW and MW frame aerial coils are intact, as are the two windings on the oscillator coil. The wave-change switch contacts have been cleaned with Deoxit, and the continuity of the contacts seems fine. The valve-holder contacts look intact and I've cleaned them with Deoxit and a 'teepee' orthodontic brush.

Silver mica caps are generally very reliable but I have heard of the rivets working loose and the caps failing on DAC90As. There are four in the front end, and to disconnect one end to test them makes them unsuitable to re-solder, so I obtained 1% replacements from Hi-Fi Collective and fitted them. (on test, the removed ones seemed not to be faulty). Of the five resistors, R2, R4 and R5 were more than 10% out (spec is +/- 20%), so I replaced them just in case.

All the valves work fine when placed in the working set, and when the working set valves are fitted in the non worker, it still fails to work.

The fact that an IF signal injected into the control grid of V1 produces a result seems to suggest that the pentode section of the valve is working, so perhaps point to the triode circuitry being inactive. As I have a working DAC90A, I’ve placed it adjacent to the non worker, and with the worker tuned to R4 on LW, and the non-worker tuned across the LW band, there is a strong heterodyne, which suggests the non-worker oscillator is working on LW. When on LW, both the LW and MW fame aerials are in series which confirms that both coils are intact. However, when the LW coil is switched out on MW and the non-worker tuned across MW band, no heterodyne is heard on the working set, so seemingly no oscillations on MW. The trimmers and oscillator coil are all sealed as they were at the factory, so I’m loathe to try to realign what - for now - is a non-working front end.

Ordinarily, I’m not a fan of ‘lucky dip servicing’ – just changing components willy nilly, hoping that one or another might be the culprit. Hence, I’d tried to adopt proper diagnostic processes. I’ve checked the voltages (which the spec say should be done on MW not tuned to a station) and have compared them with the spec and with the working set. The mains voltage here is 237 Volts, so I have the non-worker set on that tapping rather than 250V. As found, the working set was on 250V and worked fine. Clearly that’s kinder to the valves and other components, but for a more meaningful comparison, I also put the worker on the 230V setting.

I’ve never had a DAC90A with all voltages exactly as shown in the spec, and though I’ve used a digital multi-meter rather than the AVO 7 used in the spec, I doubt that’s had much influence on the readings. Nothing leaps off the page to me which suggests an issue.

Sorry it’s such a drawn out diatribe, but if anyone has trodden this road and arrived at a successful conclusion, I’d really appreciate some pointers to help restore my sanity.

[cathoderay57]

Hi David, referring to Trader Sheet #1161, I'm assuming you've already replaced the wax paper capacitor C10 (0.05uF) the oscillator anode and IF screen decoupler? That being the case try replacing C21 (16uF) the second HT electrolytic. In a couple of cases I've found that even though the set doesn't hum the second smoother can have lost enough capacitance such that the oscillator fails to operate. Cheers, Jerry

[ms660]

I would check the oscillator using a 'scope.

Lawrence.

[Andrew2]

Hi Dave. That's an interesting post. I'm a little confused by your long wave results. With the worker tuned to 198 kHz and the non-worker tuned across the LW band, I can't see how you should hear a heterodyne on 198 kHz. The LO in the non-worker should be tuning over the range 615 to 715 kHz (approx) so there should be no interference to Radio 4.
This may be a pointer to the LO being massively off-twiz, which may be a further pointer to the lack of results on MW. If you have a 'scope there, it should be possible to see just what's going on.

[Ian - G4JQT]

Yes, I'd certainly investigate the frequency of the LO being 465kHz above the dial frequency. So that's the range 996 to 2070 kHz approx for MW, and 640 to 745 kHz for LW. Maybe something amiss with L1/L2?

Ian

[slidertogrid]

I am sure I read somewhere that someone had a problem with the core falling apart in the IF transformer the ferrite part coming loose from the adjuster I think.. Never had the problem myself but may be worth checking?
Rich

[Ambientnoise]

I had one which foxed me for ages, it was a shorted trimmer cap. Not easy to find as a coil is normally across them and dc conditions are usually unaltered. Only way is to disconnect and measure, preferably as the trimmer is adjusted as the short was only there in a specific place. Presumably a metal whisker but I never found it.

[Ian - G4JQT]

Yes, I've had one where the IF core came away from the brass threaded adjuster.

I've also had one that went VERY crackly at random. Touching anything inside the chassis set it off. That too turned out to be a shorting trimmer cap.

[Graham G3ZVT]

Originally Posted by Ian - G4JQT

Yes, I'd certainly investigate the frequency of the LO being 465kHz above the dial frequency. So that's the range 996 to 2070 kHz approx for MW, and 640 to 745 kHz for LW. Maybe something amiss with L1/L2?

Ian

I agree, also if you have a general coverage communications receiver (I imagine you do David), it can be more useful than a frequency counter for a task like this.

[Davewantsone]

Have you checked the valve socket for good connection to the valve?

[David G4EBT]

Thanks for reading my wordy post and for all the helpful comments guys.

Firstly, regarding the heterodyne, it's my (mis?)understanding that that if two adjacent radios with the same IF frequency are tuned to the same station and the local oscillator is working in the non-working radio, a heterodyne should be heard on the working set. That said, as the chassis of both radios are outside their cabinets and inverted on cradles I didn't check the point on the band at which the non-worker produced a heterodyne in the worker.

As to possible problems with the IF cores, given that both IFT1 & IFT2 are spot on 465 kHz when a signal is injected into the control grids of V1 & V2, suggests that the IFT cores are OK.

The resistance readings of both the frame aerial coils and the two windings on the oscillator coil are in line with the Bush datasheet.

It's been suggested that one or another of the four trimmers could have a short, and admittedly they're the only front-end components not checked as it's such a faff to get at them to unsolder the connections, (neither have I checked the twin gang tuning capacitor for shorts). I'm coming to the reluctant conclusion that's is about the only thing left to do.

As highlighted on the circuit in my first post, all silver mica caps in the front end (baring the two in the 1st IF) have been changed, as have the paper caps C1, C4, C9 & C10.

I've heard it said that the valve-holders can be suspect, so I looked into the V1 valve socket with an illuminated magnifying glass and it looks fine. As I said earlier, I cleaned the sockets of the holder with a 1mm 'teepee' orthodontic bristle brush and Deoxit.

Ironically, I've probably spent more time on this than had I stripped the chassis bare and done a total re-build al la Kalee20's magnificent total rebuild!:

https://www.vintage-radio.net/forum/...ad.php?t=35784

Crikey, twelve years have flown by since then, which is rather scary!

Not directly relevant to the thread, I've done the 'pin 4' mod on the V4 valve-holder, which produced a worthwhile reduction in hum.

I also noted in several posts concerning the Celestion speaker in DAC90As, that as often as not the speaker cone becomes detached from the bellows which can cause the speaker to sound 'raspy' or 'gritty'. I checked the speakers on these two radios and yes, the cones were indeed detached. As others have done, I eased the cone from the bellows with a little wooden throwaway coffee stirrer, carefully applied glue with a cocktail stick in the gap around the perimeter of the bellows and weighted the cone with a roll of solder till the glue had set. Sure enough, the slight 'gritty' sound has disappeared.

I'll keep bimbling along with it, (more in hope than expectation!) and report back on any progress.

Thanks once more the the helpful suggestions.

[Simon Gittins]

Originally Posted by David G4EBT

Firstly, regarding the heterodyne, it's my (mis?)understanding that that if two adjacent radios with the same IF frequency are tuned to the same station and the local oscillator is working in the non-working radio, a heterodyne should be heard on the working set.

Not quite, the radios won't be tuned to the same station but spaced apart by the IF frequency. With the working set tuned to a station at the high frequency end of the band, tuning the faulty one through the band should result in a whistle.


Looking forward to finding out what the cause of the problem is!

[Sideband]

Originally Posted by Ambientnoise

I had one which foxed me for ages, it was a shorted trimmer cap. Not easy to find as a coil is normally across them and dc conditions are usually unaltered. Only way is to disconnect and measure, preferably as the trimmer is adjusted as the short was only there in a specific place. Presumably a metal whisker but I never found it.

Yes I've had this as well....not on a Bush but on a Philips set that refused to oscillate. Just an experimental tweak of a few degrees on the oscillator trimmer and suddenly it burst into life! I didn't replace the trimmer....I decided to leave it as it was since the fractional tweak didn't really make any difference to alignment. 10 years later it's still working....

[AD360 Rob]

Morning David. As odd as it may seem, I have a DAC90A with a very similar fault although on my set, it is the oscillator that does not oscillate. Like you I have replaced (one at a time) all the capacitors in the front end, the oscillator coil, aerial coils, tuning capacitor, V1 valve base, various out of spec (but only slightly out) resistors, the waveband switch, All trimmers have been tested and then substituted. I have also tries more than a few UCH 41s. all taken from working sets. The radio has actually worked under it's own steam once for about 5 minutes. Injecting the local oscillator frequency from my Advance E2 results in a working set. This really has me tearing my hair out. I've even subbed the 1st IF transformer and dropper resistor (it was an R.S. component). It's currently on the shelf in the workshop where I can glare at it occasionally. I will substitute the main smoothing can as Cathoderay suggested in post #2 and see what that does. !

[David G4EBT]

Hi Rob,

Thanks for reading the thread and for your post.

Oh crickey,' you're further along your 'voyage of discovery' than I am, yet have still to reach your desired destination! That doesn't bode well for my set. Back when I first started on this set I replaced several caps and one or two out of spec resistors. As the residual hum level was fairly high, though reservoir/smoothing cap, on DAC90As have proved to be very reliable., I replaced it with a new BVWS sourced one. It made no difference and turned out to be the UL41 output valve that was the main culprit.

As I said in my first post, given that only the front end now seems to be the problem, with the rest of the radio seemingly working OK, one by one I replaced the silver mica caps with riveted connections and a couple of 'micamold' caps. I also replaced the front end resistors even though they were only a little out of spec. I'm quite embarrassed to say that apart from those in the IFTs, there are now no original capacitors in the set, and apart from those in the power supply, there are only four other remaining original resistors.

I know that the wave-change switch on DAC90As has often been problematical, as have the valve-holders. A bit of a faff to replace the switch with a Lorlin one and as it seems to be working fine, I've not yet tackled that. Likewise, the valve-holder seems to be fine and as you've replaced each of those with no success, I'll leave that for now.

Simply for a process of elimination, more in hope than expectation, I'll disconnect the four trimmers in turn to see if that throws up a fault and will check the tuning cap.

Within the limits of my abilities I've always tried to adopt a methodical diagnostic approach - measuring voltages, checking resistors and caps, checking the continuity of 'switchey', close visual examination, signal injection and tracing, rather than the 'SWAG technique' ('Stupid Wild Ass Guess'), or 'lucky dip' - changing parts one by one till the set hopefully bursts into life. But when I stare at the chassis now, it looks like an exemplar of both of those latter approaches.

First pic below shows the caps I replaced some years ago.
Second pic are the rivetted silver mica caps I replaced recently in the front end.
Third pic shows the underside of the chassis as it now is.
Fourth and fifth pics for the benefit of anyone who might have a wave-change switch problem, showing how to replace it.

Ho hum - sun's out and the lawns need mowing, so duty calls.

[AD360 Rob]

Well, I changed the main smoothing capacitor today (it was an RS one like the dropper) I used a good original that read almost perfect for capacitance and leakage, guess what? . . . no difference whatsoever. It's back on the shelf as I can hardly bear to look at it right now. I have checked again valve voltages and as before, the only ones that are wrong are V1 triode anode at almost double what it should be (89v instead of 47) and cathode is reading zero instead of 0.8v. Oh well, I'll leave it another month or so, plenty of other sets crying out for attention.
Rob

[John Caswell]

The fact that the cathode volts are zero implies that section of the valve is not drawing any current. Are you sure the cathode pin on the valve is actually making contact with the base socket? A method I use is to wind a thin (24 swg) piece of wire around the pin and run it up the side of the valve taping it in place - refit valve and measure continuity between wire and cathode bias components. I have had valve base pins break in the middle and the McMurdo variety lose both the "prongs"

John

[AD360 Rob]

Morning John, that's a good suggestion as I did think when I measured it that "I've had this before" (in an amplifier which turned out to be the valve having no emission) The valvebase is new but that means nothing so I will give your idea a try and see what gives.

[McMurdo]

Always disheartening I know when you've done so much, and all the right things in the right order...but think of the jubilation when you've cracked it!
I'm always suspicious of those McMurdo forked valveholders, I've had a few tussles with them which is why my username was picked in their honour. The FC position is the most fiddly to change but I wonder if it's worth swapping?

Interestingly, you can get heterodyne whistles from the working set, in many cases when the suspect set is actually switched off. I'd also use a scope, you can probably loose-couple it, and make sure the frequencies are at least in the right area.

[David G4EBT]

Thanks guys for reading the thread and for your helpful responses.

It seems that the three main shortcomings in the front end that have been mentioned in this and other threads are the valve-holder the wave-change switch and the trimmer caps. Yesterday, I removed the bank of four trimmers and they tested fine, so this mornings task will be to reinstall them, which is quite a fiddly task, but at least it's exonerated the trimmers.

The continuity of the wave-change switch seems fine. Looking into the valve-holder with a magnifying glass it looks OK, and as the valve pins are 1mm diameter, I've pushed a 1mm drill shank into each hole, and the forks grip the shank. It may however be that replacement of the holder is the only untried option to pursue.

I do have a scope and frequency counter, and an isolation transformer to power the (live chassis) radio, but I admit to being a tad reluctant to connect my scope to any high voltage points, albeit if set to AC input I know that it has an internal blocking capacitor to protect the scope from a high DC voltage. It also has a 20V/Div setting. At present, the scope is working but the radio is not, so I don't want to end up with two restoration projects! Oh well, I suppose it's a case of 'no guts - no glory' so I'd better pull myself together. .

I guess I could use an RF sniffer adjacent to the oscillator coil with my frequency counter.

The voltages around V1 don't look out of the way compared to the spec and to my working DAC90A:

Anode: Spec is 98V, non-worker 82V, worker 85V
Osc anode: Spec: 47, non-worker 42, worker 51
Screen: Spec 47, non-worker 42, worker 49
Cathode: Spec 0.8V, non-worker 0.84, worker, 0.9.

This set is beginning to evoke the feelings of John Cleese in Fawlty Towers when he gave his car a last warning before thrashing in with a branch torn from a tree.