Heathkit RF1U - Rectifier replacement & 'tweaks'

[David G4EBT]

Back in 2018 I wrote up my experiences and observations of restoring a Heathkit RF1U signal generator in the Autumn edition of the BVWS edition (Vol43, Issue 3). However, not all forum members are members of the BVWS so some of the aspects I covered might be worth repeating.

One well known shortcoming of this, (and other Heathkit test gear which uses the Sentercel C2D selenium rectifier), is that with the passage of time, it fails to give the required DC output. The input to the rectifier from the mains transformer is 150V AC, which - if the rectifier is in good health - should provide 160V DC. That is then smoothed by three 20uF reservoir/smoothing capacitors in one can and a 2K2 load resistor across the caps to provide 150V HT. Clearly if the HT is appreciably lower, the generator will not perform as well as it should, and may not work at all.

It can be replaced by a 1N4007 silicon rectifier, which will provide too high a voltage so will need a series resistor to bring the voltage down to 150V. I found by trying several values that a 560 Ohm resistor did the trick. As to how to mount the new rectifier and resistor, it seemed to me that rather than leave the components 'dangling', it would make a neater and safer job to make a little bracket and turret board to mount on the same chassis holes as the Sentercel.

The three test sockets (‘AF In, AF Out & RF Out’) are TV type connectors, known in Europe, Australia, and maybe elsewhere, as ‘Belling-Lee connectors’ (‘IEC 61169-2 radio-frequency coaxial connector of type 9,52’). The Heathkit manual give instructions on how to make up a test lead to fit the sockets. They’re perfectly functional, neat, and original. Given that Belling Lee sockets and plugs are widely used for terrestrial VHF/UHF roof antennas, antenna signal amplifiers, CATV distribution equipment, TV sets and FM / DAB-radio receivers, they’re well up to the task, so until recently, I’ve left them in situ.

However, for many years, most test gear has featured ‘BNC’ connectors and it was becoming increasingly inconvenient to use the Belling Lee sockets, so I decided the time had come to replace them with BNC, which no doubt other users have done long before me. I used the type of BNC sockets which have a square flange with a hole at each corner for chassis mounting. Luckily, two holes (diagonally), line up with the mounting holes of the old sockets and the screws and nuts can be re-used. However, as the old sockets were recessed and the BNC ones are surface mounted, I had to extend the wires of the disc capacitors which connect to the centre pins of the AF sockets. I did that with little wire 'spirals', wound around a 0.8mm drill bit. (The RF socket doesn't have a disc capacitor and there is enough slack in the wire to reach the new socket).

I've also added an additional BNC socket on the rear of the RF1U so I can directly connect my frequency counter - a useful addition which another forum member mentioned that he done. I'll cover that in another post.

I hope these experiences might come in useful for other RF1U owners.

Pic 1: A sketch showing the circuit of the rectifier section and the replacement rectifier.
Pic 2: The old Sentercel C2D alongside the replacement.
Pic 3: The front panel with the original Belling Lee sockets in place.
Pic 4: The new BNC socket fitted.
Pic 5: The rear of the AF BNC sockets, with the wires from the disc capacitors (C20 & C22) extended to reach the centre pins.

[Freya]

A great article David, made for good reading.
One thing that crossed my mind is if you could have mounted the diode across the Sentercel rectifier, thus not needing the resistor (would only be viable if it was low voltage, and not let out the stink). I do the same thing with the finned selenium type rectifier in televisions, but not yet tried with a Sentercel.

[David G4EBT]

Often, when wanting to monitor the frequency of the RF1U with a frequency counter, if the attenuator is set to a low level of signal, there may be insufficient drive to operate the counter, which can sometimes be inconvenient. Another forum member mentioned that he'd overcome the problem by fitting an additional BNC socket mounted on the rear apron of the chassis next to the mains input. The socket is connected to the RF output prior to the 'fine' control of the step attenuator so it received the full unattenuated RF output.

The 'fine' control potentiometer is mounted inside the screening box which covers V1, the attenuator and coil turret, so it is more convenient to connect the BNC socket after C25, which connects to pin 6 of V2, which is outside of the screening box. The wiring around V2 is a bit of a 'rat's nest' but I hope the annotated part of the Heathkit diagram which I've marked in blue makes that clear.

The forum member who brought this modification to my attention also mentioned that if it's desired to power one of those small frequency counter displays which abound on internet, a small bridge rectifier (EG WO1) could be mounted under the chassis and connected to the heater supply to 'rob' some power to provide a DC supply to the counter. I didn't add that feature as my counter is powered separately, but for completeness, I've included it on the circuit as other users might wish to add it.

I used a step drill to drill the hole for the BNC socket on the rear panel of the steel chassis, which it did in a jiffy with no centre punch, no pilot hole and no burrs.

Pic 1: Circuit add-one for counter, and if desired, a low voltage DC supply to a small counter display.
Pic 2: Where to connect the BNC counter socket on the tag-strip adjacent to V2. on the output side of C25.
Pic 3: BNC socket fitted to rear apron if the chassis adjacent to the mains flex power input.

Hope that's of interest.

[wd40addict]

Stephen: If the diode is in parallel with the selenium rectifier as a 'helper' you still need the resistor. Or is this not what you're proposing?

[Valvewhiz1]

Hi David,

I just dug out my RF-1U and added the BNC to the back to drive my ancient YC-355D counter.
It’s amazing how stable these little generators are once they are warm.

Regards
Amanda M0DZO

[David G4EBT]

Thanks for reading the thread and for your comments Amanda.

Yes, there's very little drift. Probably helped by the coil turret being beneath the chassis in a screened box. V1, (which has a screen), is sited close to the tuning capacitor, but the V2 is well away and there isn't much else above the chassis. On the two that I've restored, the dial accuracy was for the most part within spec, and setting the pointer on an analogue dial isn't in itself that accurate when if the pointer is one side of the line or another, it's enough to shift the frequency. I've attached two tables which show the accuracy of two RF1Us after restoration.

For anyone who doesn't have a frequency counter, the small amount of calibration that's called for doesn't need any instruments and is outlined in the manual, a full copy of which is now in the Heathkit archive on the forum.

The third attachment shows the specified voltages showing the results on the two that I've restored, and is annotated with notes about how the voltages will vary on some settings, as stated in the manual.

The fourth attachment is of another early version which I bought at a rally in late 2019, which awaits restoration. As is often the case, the original Heathkit knobs have succumbed to the ravages of time, and the makeshift assorted knobs suggest that whoever the last owner was, they weren't very fussy custodians, which might not bode well!

The last pic shows the hazardous original twin core single insulated mains wiring which is connected to a tag-strip and from there, two wires goes to the mains switch on the AF control, and back again to the tag-strip where it connects to the mains filter. there's nothing to stop the mains flex twisting around which could cause one or the other joints fracturing, and the metal cabinet isn't earthed. That's very deficient, even by the standards of the day.

I'll add a few more comments and pics in another post.

[David G4EBT]

The mains RF filter is rather grand, consisting of two two pile-wound 1.1mH RF chokes and four .001 (1,000pF) ceramic disc capacitors. Some might argue that they should be replaced by Class X or Y, but it's not as though they're wax paper caps. In addition to the RF filter, there is a screening box covering the coil turret, RF attenuator components associated with V1.

Anecdotally, I believe that towards the end of their production run, which spanned many years, neither the RF filter nor the screening box were present. It makes me wonder whether the original reason these components were included, was down to meeting FCC standards on RFI. In the early version that I restored, and another early one I now have awaiting restoration, both the chassis and the steel screening box are copper plated, which is rather fancy!

Given that most hobbyist test gear spends most of its time on a shelf gathering dust, the saving grace is that whilst resistors might not retain their value and may need replacement if far out of tolerance, the valves will have not had a hard life. Nor for that matter, the 'switchery' which might just need a squirt of Deoxit or Servisol.

I haven't come across an RF1U in which the triple unit (3 x 20uF) reservoir/smoothing capacitor hasn't reformed. One section is the reservoir - the other two in parallel are for smoothing. It's a small tubular can mounted above the chassis 7.5 cms x 2.5 cms (3" x 1"). Should it fail, if desired, to preserved the appearance, one 22uF and one 47uF replacement tubular capacitor would easily fit inside the can.

In conclusion from my perspective, the RF1U is a capable, reliable, compact and attractive little signal generator which is well up to the needs of hobbyists. Apart from the selenium rectifier and the safety shortcomings of the mains input arrangements, and a few resistors perhaps having drifted high, little else seems to go wrong with them, and it's well within the capacity of hobbyists to repair, restore and maintain. Helped as always with the detailed and copiously illustrated Heathkit manual, which calls for no prior experience of electronics construction or test gear.

Unlike vintage radios of unknown provenance, mercifully, the factory aligned coil turret of the RF1U doesn't have accessible cores and there's only one trimmer to adjust as instructed in the manual. Hence, an RF1U is unlikely to have received the unwanted attention of maladroit 'phantom twiddlers' but will be invaluable in aligning any vintage radios that have been dabbed with by these varmints who walk among us!

I hope these rather rambling notes might be of interest and help to someone.

Pic 1: A sketch which shows more clearly the mains input and meandering wiring to the mains switch and back.
Pic 2: The copper plated screening box.
Pic 3: Mains input filter, three core replacement mains flex and chassis earthed, failed Sentercel rectifier replaced with 1N4007 & series resistor.
Pic 4: Shows the reservoir/smoothing can above the chassis.
Pic 5: A sketch of the power supply.

[Valvewhiz1]

Hi,
Regarding the output connectors being Belling Lee tv types, I was just going to make up a test lead as I haven’t swapped the sockets on the front, then I remembered the BNC female to belling Lee male adaptors I bought reasonably cheaply off a well known Auction site, problem solved.

Amanda M0DZO

[Valvewhiz1]

Hi again,
David I have two of these, the first one I had spent some time in the loft and damp and squirrels did it no good, that was the later cream version, the knobs got chewed badly and the centre tuning shaft was broken. The other one is the earlier grey front one and is pretty much mint.

It’s fitted with a 3 core cable off a bbc computer and works well, one of my signal generator collection which includes an avo military one, Marconi TF995, Marconi military one similar to the tf995 in a massive case, a tf2015 and 16, advance e1 and e2.

Regards
Amanda M0DZO

[DonaldStott]

Hi David G4EBT - really good to have all your Heathkit RF1U modifications and improvements assembled in one place. Previously I used your detailed Thread from October 2017 along with your thorough article in Vol 43 Autumn 2018 edition of the BVWS Bulletin - both excellent sources. These helped me in the restoration of my own Heathkit RF1U - details of my amateur fumbling can be found here:-

https://www.vintage-radio.net/forum/...d.php?t=167914

As you know I have a Topward Function Generator which has a Frequency Counter function but I am not getting any readings when I connect the RF Out from the Heathkit RF1U. From this current Thread I have noted that " ... if the attenuator is set to a low level of signal, there may be insufficient drive to operate the counter..." Looking back at Post #11 of my own previous Thread I am fairly satisfied that the RF from the Heathkit RF1U is indeed working although some further minor calibration may be required?

The addition of a new BNC socket on the rear of the Heathkit RF1U seems like a fairly straightforward Modification if you feel that this would be appropriate for my Topward Function Generator?

[wd40addict]

A final post script from me on this: As we saw in the other thread the output from some of these units has poor distortion on both the internal oscillator and the modulated waveform. I set off on a little odyssey to see if this could be improved.

The output from the unmodified unit is shown in the first two pictures. The THD+N on the internal generator is 15% measured on a Heathkit IM-58u. The modulation depth of the RF output is way more than the usual 30%.

Pictures 3&4 show the output from the modified unit, THD+N is now 4%, with a mainly 2nd harmonic residue and the modulation depth is now around 30%. Modulation frequency has fallen from 457Hz to 340HZ, I haven't bothered to try and tune for the textbook 400Hz.

The modification is to remove the transformer yellow wire from pin 8 of V2A, then reconnect it via a 470R resistor in series. C14 also connects to pin 8, but should be left in situ, see modified circuit. R8 was left at 100k.

Now this works fine on my sample of 1, but try at your own risk. Easily reversible if required.

[David G4EBT]

Originally Posted by DonaldStott

As you know I have a Topward Function Generator which has a Frequency Counter function but I am not getting any readings when I connect the RF Out from the Heathkit RF1U. From this current Thread I have noted that " ... if the attenuator is set to a low level of signal, there may be insufficient drive to operate the counter..." Looking back at Post #11 of my own previous Thread I am fairly satisfied that the RF from the Heathkit RF1U is indeed working although some further minor calibration may be required?

The addition of a new BNC socket on the rear of the Heathkit RF1U seems like a fairly straightforward Modification if you feel that this would be appropriate for my Topward Function Generator?

It depends on the minimum signal level that your frequency counter calls for to operate. The spec states that the RF output on all ranges is up to 100mV, and if your RF1U attenuator is on the maximum 'Coarse' setting, that's what it should provide. If you then use to 'coarse' attenuator steps to reduce the signal level, at some point the counter may cease to register the signal.

When aligning a radio, the generator mimics the very low RF signal levels that the radio will receive if working correctly and properly aligned, so the aim is to use the lowest output from the generator to achieve that objective. However, on the lower settings, depending on the sensitivity of the counter, the signal level may be too low. By adding the BNC socket connected prior to the attenuator, the counter will receive the full 100mV output of the RF1U. (The spec states that the maximum AF output is 9V p-p).

Hope that helps.

(When you say that 'you're satisfied that the generator is working', what does the unmodulated and modulated output look like on your 'scope?).

[David G4EBT]

Originally Posted by wd40addict

A final post script from me on this: As we saw in the other thread the output from some of these units has poor distortion on both the internal oscillator and the modulated waveform. I set off on a little odyssey to see if this could be improved.

The modification is to remove the transformer yellow wire from pin 8 of V2A, then reconnect it via a 470R resistor in series. C14 also connects to pin 8, but should be left in situ, see modified circuit. R8 was left at 100k.

Now this works fine on my sample of 1, but try at your own risk. Easily reversible if required.

That's a useful tip and an easy modification. I'll bear it in mind when I get around to restoring the one that I've yet to get on the bench.

Given how many seem still to be around, quite a lot must have been sold over the many years the kit was in production. No doubt there may even be some that were never started, or never finished. The serial number on the one to which I've just fitted the extra BNC socket is: 3095-8, scrawled on in ballpoint pen which is meaningless unless we know what that number conveys. '3095' could mean the number of kits to that date, '8' could mean eighth month, or [196]8 or [197]8, or none of those. I guess we'll never know.

All mine have been rally acquisitions in a bit of a sorry state.

[wd40addict]

The RF-1u seems to have been one of Heathkit UK's best sellers. Mine has the beige finish and 1969 capacitors, but doesn't have the 1970 R8 errata. So was presumably built in 1969. Can't remember where I got it, probably a rally 20+ years ago when no-one was particularly interested in Heathkit equipment.

The screening is not for FCC, but to try and stop spurious RF leaking into the radio under test. The RF-1u attenuator could be set to produce outputs in the microvolts and in these circumstances you need to ensure that the only path into the DUT is via the coax cable and not radiated through the box or the mains lead (hence the mains filter).

Attached is the sideband output of my modified generator. As mentioned by others above frequency stability is pretty good for such a simple bit of kit.

[DonaldStott]

Thanks David - as you predicted "If you then use to 'coarse' attenuator steps to reduce the signal level, at some point the counter may cease to register the signal." With the Heathkit RF1U dial set at around the 465Kc mark and with the FINE and COARSE controls at maximum this figure is indeed displayed on the Topward. When I turn the COARSE control down to '3' the Topward counter ceases to register the signal - only visible at COARSE setting '4' and '5'.

So it would appear that I may benefit from the addition of a new BNC socket on the rear of the Heathkit RF1U and the internal modification?

[David G4EBT]

Originally Posted by DonaldStott

Thanks David - as you predicted "If you then use to 'coarse' attenuator steps to reduce the signal level, at some point the counter may cease to register the signal." With the Heathkit RF1U dial set at around the 465Kc mark and with the FINE and COARSE controls at maximum this figure is indeed displayed on the Topward. When I turn the COARSE control down to '3' the Topward counter ceases to register the signal - only visible at COARSE setting '4' and '5'.

So it would appear that I may benefit from the addition of a new BNC socket on the rear of the Heathkit RF1U and the internal modification?

I think so.

You need to be careful when drilling the hole in the chassis.

If you don't already have a step drill, it's worth getting one (or a set) with which you can quickly drill the hole in the rear of the chassis with no faffing about centre punching or pilot hole to drill, no snatching and no burrs to remove. You just go down a step at a time till you reach the desired size. Excellent for acrylic too - no splintering or risk of cracks due to snatching.

https://www.youtube.com/watch?v=cVs3tLZQZ6g

You don't need an expensive one - they get used so little they seem to last forever:

https://www.ebay.co.uk/i/30380653181...nd%3DUnbranded

Despite them often being called 'cone drills' they're not. Cone drill aren't stepped - they're conical drills.

If when drilling a hole in sheet material it's slightly too small but the next size up will be slightly too large, a reamer is useful to have around:

https://www.ebay.co.uk/itm/124247215...oaAnDkEALw_wcB

Those are just examples from UK suppliers - I'm not recommending them - you can search for 'step drill' and 'reamer' and make your own choice.

Hope that helps.

[DonaldStott]

Thanks David - fortunately I have a set of step drills and have, so far, never needed to resort to using a reamer!

Looking at your pictures above I note that the additional BNC socket is fitted to the rear apron of the chassis adjacent to the mains flex power input but I'm not sure how this BNC socket has been secured?

[David G4EBT]

Originally Posted by DonaldStott

Looking at your pictures above I note that the additional BNC socket is fitted to the rear apron of the chassis adjacent to the mains flex power input but I'm not sure how this BNC socket has been secured?

I used one with a locking nut and washer rather than a square one with a hole at each corner (the type which I used on the front panel to replace the Belling Lee sockets).

[DonaldStott]

Originally Posted by David G4EBT

I used one with a locking nut and washer rather than a square one with a hole at each corner

Thanks - Order placed.