6J7 versus 6K7 in a CR100

[RogerWalker]

Can anyone explain why the CR100 manual shows both 6J7 and 6K7 as equivalents for KTW61? I know the 6J7 as a 'straight' audio amplifier, a bit like an EF37, whereas the KTW61/2 and the 6K7 are all variable mu devices. Surely, the AGC wouldn't work properly with a 6J7 in the RF or IF stages?

I have read the long CR100 threads on here, but couldn't find any mention. I have also read the excellent CR100/1 and CR100/2 refurb stories, which mention fitting a 6J7 in the BFO, with some ill-effects at high frequencies.

I am interested because I am working myself up to refurbishing a CR100 during this winter - this one has two of the KTWs replaced by metal single-ended 6SG7s in the IF strip - at least the 6SG7 is also variable mu, so should work OK, albeit with a bit more oomph, having higher GM.

I am dreading having to restring the drive cord and to replace 32 0.1uF decoupling capacitors - but the exercise putting it on the bench and flipping it over will do me good!

[Andrew Sinclair]

The CR100 manual was written for people trying to maintain these sets, many were fitted to Royal Navy vessels and could see service around the world. They might not have the correct valve in the ship's stores, but needed to keep the set working. The KTW62 is close to a 6K7 but has higher gain. I think the 6J7 and 6K7 were suggested as replacements to keep the set going if the proper valves were unavailable. My copy of the manual (T1868, Aug 1945) suggests they should only be used in an emergency.

These days, with many used valves in all sorts of conditions, it probably doesn't matter too much. There are early KTW62s in bottle shaped envelopes and later ones in straight envelopes. I think the electrical differences between them might be greater than the difference between a 6K7 and an early KTW62
The 6SG7 is a much "hotter" valve and might cause problems. I would advise using KTW62s if you can find them.

Watch out for poor insulation in the tag strips and system switch wafer. On mine, the leakage currents between the HT rail and AGC line were bad enough to cause a positive bias on the AGC rail of a few volts, which ruined the AGC performance.
I fitted a ceramic wafer for the system switch and made new tag strips.
These are interesting sets and saw use in all three services. The RN used them into the 1950s at least.

By the way, I don't think any version of the CR100 was fitted with an S meter when it left the factory, despite what many websites claim. It was just a very common amateur modification.

[RogerWalker]

Thank you - I guess needs must during wartime.
The 6SG7s are [I now notice] fitted in the RF section and have different pin connections, being single ended. At this stage, I am not in a hurry to buy more KTW62s and retrofit them - but if the receiver turns out to be insanely unstable, I guess I may have to.
Thanks for the tip about tag strips - they are something one assumes will be OK unless tracked across.
Mine doesn't have an S meter fitted but it does have a magic eye tuning indicator with a diecast aluminium escutcheon which I must say looks pretty original. It also has an EA50 fitted in a noise limiter circuit that again looks like it is original - maybe EB34s were in short supply when it was built?
The model/version number plate is missing from the front panel so I am struggling to determine which version it is. Going by the manual and the rear panel, it looks like it may be a 100/4 - but it doesn't quite tally with the manual. No doubt I will get a better idea once I have its internals disgorged like a patient being operated on in Casualy or Holby (less the tomato ketchup, hopefully)! I was rather hoping that the Scots person I bought it off in 2013 might recognise it from my description, as he might know more about its history. I will live in hope that he is a member here.
Thanks again for your valuable insights Andrew.

[turretslug]

As said, I think that the equivalents listing reflects the wartime maritime situation, i.e. "this will at least get it going". ISTR that some equivalents were shown in brackets where they would be less-than-optimal but would still be functional. The early-war atmosphere is further reflected where it says not to discard noisy or low-emission valves but to re-deploy them as appropriate to their shortcoming, i.e. low-emission valves in early signal stages and noisy ones in later stages. I suspect that the lower gm of the 6K7 would be less noticeable in the IF stages than at the higher RF stage frequencies, there are 3 of them in the IF after all! Interesting that the 6SG7s were fitted in the IF, rather than RF stages- there was all sorts of urban wisdom about hotting-up the front ends of older sets with different valves but whether much attention was paid to optimising operating conditions, signal-handling and AGC characteristics is debatable. I found when substituting (out of interest rather than necessity) that the LO was the most obviously critical- all KTW62s and 6K7Gs tried would happily oscillate at the top end of Band 6 (30MHz) but no 6K7G would maintain oscillation tuning below around 13MHz, sometimes petering out below 15MHz whereas even long serving KTW62s oscillated healthily right to the low end of this band at 11MHz. Obviously reflecting the somewhat lower gm of the 6K7G combining with the soggy L/C ratio at the low end of Band 6.

Incidentally, I'm not sure what the actual difference between a KTW61 and KTW62 is- I have seen the latter listed with slightly higher heater current but this may simply be a listing error (not unusual with various valve data sources....) and have also heard that it was simply a KTW61 selected for lower noise at the higher HF frequencies- a single-source anecdotal quote though. I've certainly seen several of them with KTW61 printed on one side and KTW62 on the other.

One of my idle thought-projects (i.e. I've thought of it but am unlikely to get round to it!) is matching radios to contemporary valve series that would have suited rather well but were out of the question due to marketing/commercial or political/national reasons- I thought the CR100 with the B8B/loctal (introduced 1938) series- 7H7, 7S7, 7B7 (IF),7C6, 7C5 ought to work pretty well, similarly R1155 with Y8A Stahlröhren series- EF11, ECH11 etc. Probably sacrilege nowadays, but when these sets were plentiful one might have got away with it!

[Radio Wrangler]

Sticking in more exotic RF amplifier valves was a national sport back in the 50s and 60s.

The 6SG7 offered more Gm and so more gain, but it also gave you duplicated cathode connections to halve the inductance of the cathode connection.

The KTW62 has only a single cathode connection, but it does keep the g1 connection on the top cap well away from the anode pin, reducing stray capacitance. If going to a 6SG7, it's worth duplicating the cathode decoupler on the second cathode pin. This not only acts to increase gain at higher frequencies, it also helps stability. The reputation of the 6SG7 for poor stability may come from people just shoving them in and hoping for wonders.

The AR88 came ready with 6SG7s, five of them. Back in the day, one popular hotting-up trick was to put a Western Electric 717A in the first RF position (watch out for its lower anode and g2 volts rating) It certainly looks more exotic! This is also 4mA/V like the 6SG7 and has dual cathode pins. The internal construction has the structure horizontal for even shorter connections. This is only of value if the under-chassis connections of the receiver are very short as well.

The successor to the 717A was the 6AK5 uhf pentode. Despite 7 pins, it has dual cathode pins at the expense of an internal k to g3 connection. This valve pushed Gm up to 5.1mA/V. Anode ratings are similar to the 717A. The connections are very short due to the arrival of all-glass construction.

The KTW62 has outer metalisation for screening. The 6SG7 is a metal valve and thus screened. The 717A has a metal collar and the valve itself is well down inside the collar, giving partial screening. The 6AK5 is not screened and usually lives within a screening can.

Having messed around with some of these 'hotter', mre exotic valves in the sixties, I didn't experience any significant improvements to what I could hear. It seemed to be on the borderline of discernibility. Of course a 6AK5 on an adaptor of a B7G base grafted onto an octal plug is going to spoil those short connections.

David

[turretslug]

The nice thing about the AR88 front end is that the designers chose to make best use of the 6SG7's two parallel cathode connections by connecting both valve socket tags directly to chassis by short thick pieces of copper braid, the power supply being arranged to provide negative grid bias supplies with a minimum level of preset grid bias being added to by manual gain and AVC circuits. Many sets, including the CR100, adopted cathode (auto) RF stage bias with rolled-foil bypass capacitors with significant lead lengths- one could have a swanky RF valve upgrade with 28 parallel cathode pins but it all counts for nothing if the bypass cap's self-inductance and lead inductance dominate things. I can't help thinking that a modern miniature, minimum-inductance ceramic capacitor, rated at 100V-plus to allow for the popular positive-going RF gain technique, used for cathode bypassing plus a similar suitably rated screen-grid component would be an easier, quicker and cheaper upgrade to try before getting carried away with fitting more exotic valve types. Still, it's probably a timeless feature of human nature to feel that one's own whatever is that bit better than everyone else's and a 6AK5 or cascode ECC88 probably scores higher in bragging rights....

[gm0ekm cecil]

Originally Posted by RogerWalker

Can anyone explain why the CR100 manual shows both 6J7 and 6K7 as equivalents for KTW61? I know the 6J7 as a 'straight' audio amplifier, a bit like an EF37, whereas the KTW61/2 and the 6K7 are all variable mu devices. Surely, the AGC wouldn't work properly with a 6J7 in the RF or IF stages?

I have read the long CR100 threads on here, but couldn't find any mention. I have also read the excellent CR100/1 and CR100/2 refurb stories, which mention fitting a 6J7 in the BFO, with some ill-effects at high frequencies.

I am interested because I am working myself up to refurbishing a CR100 during this winter - this one has two of the KTWs replaced by metal single-ended 6SG7s in the IF strip - at least the 6SG7 is also variable mu, so should work OK, albeit with a bit more oomph, having higher GM.

I am dreading having to restring the drive cord and to replace 32 0.1uF decoupling capacitors - but the exercise putting it on the bench and flipping it over will do me good!

Valve data, I hope this helps.

[dtvmcdonald]

I've tried all sorts of tube substitutions in RF and convertor stages. This is with US tubes of course, both octal and miniature, no old 5-6-7 pin tubes.

I NEVER found a substitution for RF amplifiers that actually helped. By help I mean
generated an actual improved signal to noise ratio on an actual signal. I did find
increased gain in some cases. The only circuit changes I made were to grid bias in order to get the specced transconductance change. This was done with a series
battery and pot in the SGC line before the changed tube.

I rarely found much change in S/N using very low signals from a signal generator
through a step attenuator, so the "fake antenna noise" was just a room temperature
resistance of the correct value. But I did see some. Its just that it was always swanped ... even at 28 MHz ... by noise picked up by the antenna. Of course, that was a cheat, since it is a 37 foot long wire at a height of 1/16 inch, into an MPS8097 transistor emitter follower with an input impedance of 5000 ohms (no transmission line). That's a really super transistor.

On the other hand the various different heptode or hexode convertor tubes are very very different, as are the various triode/hexode or triode/heptode convertors.
The early ones are very very bad, the later ones better, and the 6BE6 is the clear winner over all others, including all the triode/hexodes or triode/heptodes in one bottle. I'm including S/N as well as oscillator stability and ability to work well at the
bottom of the highest frequency band.

[turretslug]

It may be simplistic, but I assumed that as I could peak the noise in my CR100 by twiddling the aerial trimmer with terminated input and its designed KTW62 RF stages that the laws of the universe were telling me that there wasn't much to be gained by messing around with more modern and allegedly fancy valves- beyond amplifying that noise a bit more.... Certainly, once an aerial was connected, the chances of getting improved performance with real-world signals seemed even slighter.

Multigrid mixers are an absorbing subject in themselves- whilst it became almost trendy to denigrate them wholesale as hopelessly noisy and therefore undesirable anachronisms, a bit of reading in the classic texts of the 1930s shows that they were actually carefully considered and designed devices that showed appreciable sophistication and effectiveness. No doubt the wartime acceleration of development, miniaturisation and precision in all aspects of electronic components added extra consistency in grid alignment and electron lensing, reducing mixer noise somewhat but previous discussions on the forum seemed to suggest that by time one and certainly two RF stages had been added in front of a multigrid mixer, its noise contribution became far less significant.

I never had much faith in the arbitrary nature of add-on signal strength meters, in many cases I suspect that the scale may as well be blank and used simply as a peak indicator. Magic eyes do have a certain charm and certainly give a good peak indication- I made a little set-top box for my CR100 with an EM84 displaying through a slot, fed with AGC and HT from the set- it only needs about 1mA- with its own heater supply. Maybe the set used in post #1 had parts from an R1155?