Tech TE-20D Repair

[tri-comp]

David,

You're in the UK so maybe you'll have a chance to pick-up a Brimar equivalent, the 13D6.
US type 6913 (GE & Sylvania) is another direct substitute, probably even harder to get.

It seems odd why they stuck a beefy dual-triode into this circuit.
ECC82/12AU7 has the same pin-out and amplification factor.
I wonder what will happen if you plug one in ?
It's capacitances are much smaller on all acc'ts but shouldn't that improve on starting oscillation ?
Of course it will throw-off calibration a tad.

[tri-comp]

OK, Ears red !

I went and checked what I actually put in store about 5+ years ago.
It turns out it wasn't a TE-20 but a TE-22 (Low Frequency Generator) AND a TE-20D.
I took the RF-gen. apart to have a look and here's what I found after making sure that the power supply electrolytic cap's are OK:

First of all I didn't want to measure anything with an original and possibly shot Matsushita 12BH7A in place.
Enter one NOS RCA 12BH7A.
Putting a scope on the oscillator output (V1, pin-7) reveals quite a big difference in output voltage. It obviously drops quite a lot as frequency goes up and in Band-E it's pretty much steady on 1Vpp from 10-35MHz.
At V1, cathode the signal looks fine with only minor distortion.
Lowering band to D makes things a lot worse.
Oscillator-output goes up and the cathode-follower starts to distort the signal badly.
The lower the band, the higher the clipping of the signal as the cathode-follower is fed with such a high input it hasn't a chance to pass it undistorted.
Altering the 1K cathode-resistor and thereby the operating point of the cathode follower doesn't improve on things.
Possibly a proper frequency dependant R-R/C filter-circuit to replace the 5pF coupling from the oscillator to the cathode-follower would reduce distortion.

About Band-F which in my TE-20D also has a dead short for an oscillator-coil, I can only say someone goofed at the factory.
You'll never get it oscillating with that in place.
Remember that the bands overlap and hence the top E-band and lower F-band should produce the same frequencies.
Not so with the short in place imho.

Then there's the business with the ECC82 as a replacement to 12BH7A.
I took a NOS Philips ECC82 and stuck it in.
It lives quite happily there, producing slightly less oscillator output and raising the output frequency (Band-E/20MHz) about 300KHz due to lower inter-electrode capacities.
You're not even able to register that on the front scale, so forget.
The lower oscillator-output actually lowers distortion on the lower bands a tad, but not much.
On the A-Band the distortion right out of the oscillator is VERY high and obviously things are not improved after passing the cathode-follower.

I suppose you get what you pay for !

The original 12BH7A was returned into service and I can't see any difference between that and the NOS RCA.
At least that's impressive

[Radio Wrangler]

You got what you paid for at back in the day prices, but professional instruments which once were unreachable are now affordable, or sometimes free, seen as boatanchors.

I see a good unit which does what it says on the can as a simple thing to use. I see a poorly designed unit which needs nursing along and mollycoddling to get any function as complex and too much trouble to bother with. Complexity isn't a simple matter of component count or even acreage of schematics. I think hassle factor is an important component. Large complicated looking things are just a group of simple things flying in formation.

I bought a Marconi 2008 for about the same number of pound notes that a hobbyist generator would have cost back in the day. It works on all its ranges. RF harmonics are reasonable. Modulation distortion is low. Level and frequency are accurate. I bought it non-working. I had to put the pointer drive wire back on a pulley!

Don't avoid good quality test gear, exploit it!

David

[JayBee66]

Well, I have got my TE-20D up to a somewhat satisfactory standard.

To boost the output of the AF I bypassed the 400K resistor on top of the PCB (not on the switch) to get it up to about 2.5V. It still runs at over 1KHz but I'll leave that for now.

The only component way off value was the 5pF ceramic between the two triodes of the 12BH7A, which measured 150pF. I'll replace that with a 3KV version.

The transformer outputs 160V rather than 120V so maybe that is causing problems with resistors. Who knows. I was going to sell it but I had a change of heart. I have to fix my father's GEC Starfinder so it might get used after all.

[Radio Wrangler]

Sounds like that 5pf capacitor has absorbed moisture. Electrical leakage will cause some types of capacitance meter to show exaggerated values, as well as the effect on the dielectric constant putting the capacitance up.

David

[JayBee66]

I got my "5pF capacitors" today but they all test in the mid 30s.

Of course, at that range, it could be the DMM that is at fault.

[Radio Wrangler]

If new 5pf (5.6 or 4.7?) parts indicate 30pf, then you're probably off the bottom end of the usable range of whatever you're using to measure capacitance, and you'll also ger several pf from even short test leads and hand-capacitance effects.

So as the original 5pf part read 150pf which is much higher than what the new 5pf parts read, then a likely explanation is that the original part is bad (probably leaky) and your capacitance meter isn't intended to go so low.

David

[JayBee66]

It's a 3KV blue blob and was advertised as exactly 5pF. I went for the higher voltage rating rather than the more usual brown ceramics rated at 50V because my transformer is outputting 160V rather than 120V. The voltages around the circuit are higher than they would normally be.

I am thinking of replacing the 2K 1W resistor just after the voltage regulator with a higher value to bring the voltages down. I believe the oscillator circuit is a VCO (voltage controlled oscillator) and bringing down the voltage might bring the AF down from just over 1KHz to closer to the advertised 400Hz.

What you say about capacitance from open test leads is correct. I notice that when open the leads read 30pF on the meter and subtracting that from 35pF measured when connected to the capacitor gives me a 5pF capacitor. I have learned something new!

Thank you.

[JayBee66]

I am beginning to think that this damned thing hasn't worked since before it left the factory.

Earlier, I tried to use it as an AM transmitter, playing some music from my amp into the external sockets with the mod switch on External Mod. Nothing received on a radio with a wire from the RF output draped over it.

I switched to Internal Mod. Not only was the 1000Hz AF tone was heard but also the music faintly mixed in.

I was in the process of switching off the signal generator when I heard the music clearly (and no 1000Hz tone) with the switch halfway between Internal Mod and Off.

So, that's tomorrow sorted then, fault finding the switch.

I wonder if my local rugby club want a metal ball to practice goal kicks with.

[David G4EBT]

Originally Posted by tri-comp

It seems odd why they stuck a beefy dual-triode into this circuit.
ECC82/12AU7 has the same pin-out and amplification factor.
I wonder what will happen if you plug one in ?

I wonder whether it was because it will oscillate at HF? (Assuming that the equipment into which it is installed has a coil for those frequencies!).

Interestingly, the description of the 12BH7A states that it was designed primarily for use as a vertical deflection amplifier in TV receivers and as a combined vertical deflection amplifier and vertical oscillator. However, these days it seems to be mostly in demand for guitar amp enthusiasts - hence the high price due to the magical qualities with which it is (allegedly) endowed. ('lush, mellow tones' etc). I've got a couple heading my way at a sensible price. As to the possibility of using an ECC82/12AU7, since writing that, you've tried one, thus answering your own question, which you've written about in a subsequent post. I've got a couple of ECC82s I'll give one a try out of curiosity.

I've not yet checked any voltages on mine - it was just gathering dust until I read this thread, which sparked my interest, so I must have a poke around with a DMM.

Strange that there are three of us in this thread with a TE-20D, and in each case there is no coil for the highest range - instead, just a shorting link. Unsurprisingly, none of them work on the highest band. Initially, in my ignorance I'd assumed that mine didn't work due to perhaps the valve not wanting to oscillate above 30 MHz and it was only when it was mentioned in this thread that I checked and noted the absence of a coil for Range 'F'. At some point I might check the upper and lower capacitances of the tuning cap and use an online calculator to find out what the inductance of a coil to cover that range would be, and wind one to try it in place of the tinned copper wire shorting link. It's only going to need to be a few turns.

It does seem odd that they left the factory like that. I can't see how a straight 50mm length of wire will have sufficient inductance for that range to ever oscillate.

In reality, for most practical purposes, given that the other five ranges cover from 120kHz to 38 MHz, that's more than adequate - certainly for my simple needs. It's an instrument for hobbyists, and that's fine for me - I'm just restoring old domestic radios - not putting rockets into space. Ordinarily, I use a Heathkit RF-1U, which employs a 12AT7 and ECF80. I recently bought a Leader 17A (solid state), about 25 years old, still in its original packing, never used. Works a treat and calibration is spot on.

[tri-comp]

In order to lower distortion on the RF output signal I made a few changes to the circuit:

Powersupply ripple to be reduced.
The 2nd. 20uF/150V removed and a 47uF/400V electrolytic cap. mounted instead.
The removed cap. paralleled with the first 20uF/150V cap. immediately following the rectifier.
It fits under the PCB.

Oscillator output voltage to be severely reduced to avoid overdriving the cathode-follower.
Mount a 470KOhm resistor in series with the 5pF ceramic coupling cap between oscillator anode and cathode-follower grid, keeping leads as short as possible
This shapes up the output signal very well, except on the lowest A-band where distortion is still quite noticeable, but still much, much better that original circuit.

Alter the RF output level control circuit to avoid changing the DC operating point of the cathode-follower.
Mount a 15nF ceramic or other capacitor between the potmeter center terminal ond the High-Output Red banana terminal as a replacement to the red wire already there.
Now changing the output level no longer shifts the operating point.

Reducing RF-ripple on the power supply to give just a little cleaner output by mounting a 10nF cap. directly between 12BH7A pin 6 and Ground. I used a Mullard yellow mustard cap rated 125V which seems to be enough.

I checked the 400Hz AF-generator and the output terminal, unloaded shows 448Hz and 27Vpp.
Loaded by 1KOhm reduces output to 60mV, quite useless.
Who would EVER construct an AF-oscillator with an internal Z of 400KOhm ?
Replacing the 400KOhm resistor by 1KOhm only slightly alters frequency.
Now a 4,7KOhm load leaves the advertised output of 8Vpp.
Frequency changes only 4Hz between no-load and a 4,7KOhm load.
Tech usermanual says AF-output is adjustable. I wonder how they suggest the user do that ?
Other than opening the instrument and replacing the 400KOhm series-resistor, which I did as described.
So, in a way they're right

[JayBee66]

Well, my "restoration" is complete.

In all it was an electrolytic and moulded capacitor re-cap. A replacement 5pF blue blob for the original brown ceramic between the triodes. And, the bypassing of the 400K resistor on top of the PCB to the right of the 6AR5 valve and below the green blob caps. I re-routed the white wire from the Mod Switch to before the resistor so that the voltage to the AF out and RF modulation was stronger.

Last night I couldn't get External Modulation to work with a nearby radio as receiver so I squirted some WD40 into the switch. This morning everything seems to be working fine. In the photos; unmodulated Band C, externally modulated Band C (courtesy of Tim Blake and his synthesizers) and then an internally modulated (1000Hz) signal on Band C.

The bottoms of bands A and B are clipped. I can live with that.

Job done. Thanks all.

[David G4EBT]

Very interesting and useful posts - thanks for those!

No more thoughts about any mods to try to get band 'F' working guys?

[Radio Wrangler]

I think that for Band F the amount of inductance needed to go with the tuning capacitor is simply the wiring to the switch and that direct link between switch lugs.

I also think that on this band the L/C ratio is getting so extreme that high demands are being placed on the gain of the valve, and the whole thing is very iffy.

Manufacturers of lab-grade sig gens had to go to significant lengths to get reliable VHF coverage and usually made entirely separate models for those frequencies.

I suspect that very few of these generators work at all on that frequency range.

David

[JayBee66]

Originally Posted by David G4EBT

No more thoughts about any mods to try to get band 'F' working guys?

I'm pretty sure that my Band F is working. However, I cannot show you a photo as my scope only goes up to 3MHz.

I know it is working because I get a narrow fuzzy band for Bands D & E, which are above 3MHz and the same again for Band F. No flat line.

[tri-comp]

I forgot to get my Frequency Counter from storage this last week-end.
I'll try and settle for sure if there's any oscillation taking place with the 'F'-Band setting.
I still think there's not.
My DSO is flat to 25MHz and I see nothing when the TE-20D is set to 'F'.
Set to 'E' I can still easily make out the top-most frequency at around 39.2MHz, counted by the internal DSO freq. counter.
Checking the scale calibration for 'E' at 30MHz it's spot on with the DSO !
Since the 'E' and 'F' bands overlap, the lower end of 'F' reaches well below 40MHz and should be detectable to my DSO but obviously it's not, as I still believe we're one coil short.

So, because of nothing better to do this evening I set out on trying to improve on the very badly implemented amplitude modulation of the RF output.
The 6AR5 anode-resistor was replaced with a driver-transformer from a very old transistor-radio from the times when such transformers were in fashion.
The primary winding paralleled with a 100nF cap. proved to be just the perfect replacement inductance for the 10K resistor, dropping the oscillation frequency to exactly 400Hz.
Output, unloaded voltage around 70Vpp across the full secondary winding, not using the center-tap for anything.
Now, lifting 12BH7A, pin 6 from the PCB and attaching it to a RF-choke in series with the transformer secondary attaching the opposite end of the secondary winding to the B+ where the 12BH7A anode used to be connected, should in theory produce AM at the anode. Especially when temporarily decoupling the 12BH7A, pin 8 with 1uF and disconnecting the original AF 0,1uF coupling cap. from 6AR5 anode.
Well, there is an improvement, possibly, maybe, ..I think.
Nothing worth writing about so I'll leave the poor generator with it's original setup and finally assemble it.
...until, of course someone comes up with an improvement of some sort. Possibly a suggestion for a new 'F'-coil

One final change, that really IS an improvement, and a VERY big one at that.
You can improve tremendously on the S/N-ratio of the RF signal by fixing a thick, single strand copper-wire from the PCB, where the variable capacitor is grounded soldering the opposite end to the RF-ATT potentiometer housing.

[JayBee66]

That is some good research tri-comp.

I have ordered one of those cheap frequency counter kits from you know where and I will use it to test for activity above the range of my scope.

There is plenty of room for improvement in the generator. I used it as a transmitter to a nearby AM radio and I was hearing the test signal on various parts of the dial when transmitting on Band A so I have harmonics on that band too.

Yesterday I converted my generator to three core mains cable and today I think I will do your grounding mod for improved signal/noise ratio.

Not being a Heathkit I have no qualms about hacking on this equipment. My scope and electronic switch, on the other hand, are no go areas, except for re-caps.

[JayBee66]

Originally Posted by JayBee66

That is some good research tri-comp.

I have ordered one of those cheap frequency counter kits from you know where...

which I received and it doesn't work.

That'll teach me.

[tri-comp]

Off topic,
Actually most of what I received from -you-know-where- actually works quite well.
One digital clock kit, that I paid US $2.31 for including P&P, doesn't work because of a shot 12MHz XTAL.
So, I went and ordered 10 new XTAL's for a total of Canadian $ 1.29 including P&P.
This is becomming a crazy world !!

What's with the Frequency Counter ? Can't it be fixed ( ..under a couple of $'s) ?

[JayBee66]

It was one of those cheap PIC based counters for a fiver. Either the PIC never worked or it got zapped somehow.

Still, the day has ended well when someone in the US put a Heathkit IM-2410 Frequency Counter on eBay for just £20. I snapped it up. Another Heathkit for the collection.

Even if it doesn't work it is all TTL based logic in IC packages and no pre-programmed micro-controller or micro-processor to worry about. I have already downloaded the assembly manual.