I'm not sure how to interpret the results. I presume you mean there is a typical tapped coil on the cathode and a separate 150uH choke plate load? And was that at 1MHz?

Anyway, I was thinking off the top of my head and kind of jumbled it in the text. I don't doubt that more gain increases amplitude, and the part about bias shift was wondering how much, but my first thought was on your comment about "pentode" having "higher gain." The 6BK7 osc has the plate fixed by C7 so there's not really any 'triode' to it (a fixed triode plate resembles a pentode's fixed screen) and "all the action," so to speak, is cathode current and the grid. That's why I thought being a pentode wouldn't make much, if any, difference, assuming all else were equal (which, unfortunately, usually isn't), and why I figure gm is 'the gain' in this case. Of course, tie the pentode screen to plate and you've got a triode anyway but I was thinking we could get some 'amplitude control' by lowering screen to whatever it takes. On the other hand, it's probably simpler to triode mode the thing and, if needed, simply divide down the cathode signal.
I'd think it depends on the plate load, which can be made as small as one needs to get a small signal.

Frankly, being a plagiarist, I figured on the same takeoff point as the original and, if a different tube has 20Vpp on the cathode, stick a 47k in series and, since the grid leak is 47k, that would cut it in half. Or pick it for whatever amplitude one wants.

AH HAH! Yeah, I can see that bias would have an effect. Thanks.

On the 'one buck' tube 'dollar days' theme, ABCVacuumtubes and VacuumTubes.net have the 7AK7 loctal version of the 6888 for 1 buck each. Loctal sockets are only 2 bucks at ABCV.

Ferrite isnt particularly frequency stable in a tuned circuit and a powdered iron 2 mix works well at the BCB high end at least.

Id opt for powdered iron in a High C circuit for best VFO stability if using a toroid.

Carl

Okedoke, I did a little breadboarding with a 6CM8, another 'dollar days' tube with a mu=100 triode and pentode, to see about using the pentode for osc. I used a different transformer setup so B+ is high but it proves the concept.As expected osc amplitude can be controlled with screen volts so the RC on the 6BK7 plate is moved to the screen, with a larger R of course, and the plate absorbs the B+ drop there (meaning it takes no 'extra' parts). Should work the same on 155V B+ with the screen resistor adjusted to get the same V. C3 doesn't need to be that large but I had one and that was enough to handle whatever R I came up with during experiments.

Btw, configured like the 6BK7 version I saw significant distortion during grid clamping (whether the 6CM8 was triode or pentode mode) and the 1k smoothes that out. It does distort the cathode, though (peaks on grid drive), so the take off point is under it, or on the other side of the coil if you want lots of volts. Make R1 smaller: current goes up and so does amplitude.

I learned another very important lesson: When building a circuit it helps quite a bit to plug in the right blooming tube as first time around I blind eye stuck a 6CM7 in the socket instead of the 6CM8 Don't do that. It pretty much shorts B+ to ground because plate and cathode come out the same but grid is where screen should be, which isn't all that great for the 6CM7 either.

I've tweaked a bit on the 6BK7 osc, plate B+ is now supplied through a 150uh choke which cleaned up the sine nicely(it did have a ugly harmonic in it prev)... Also there is a 39 ohm on the ground end of the osc coil that I'd forgotten about, gives a slight increase in PP voltage reading... I have no idea of this coil's ancestry, found it amongst stuff from a friends long closed TV shop... Originally I had one from a Japanese AM/FM radio, but it got changed out early on due to it was difficult to adj...

I want to wind one that could be easily duplicated(no it won't be on a 1V2) so it would give "expected" results using a AA5 or similar tuning cap... I'm going to shoot for something that will perform at the upper half of the dial, from roughly 1150-1600...

Tom

I used the AES coil because I had one and at least they're readily available, but they're not 'cheap'.

I also found the cathode resistor increases amplitude. It seems counter intuitive if one is used to Class A amplifiers, since you'd expect that to reduce bias, but, within reason, it doesn't here because bias is set by the grid clamping positive sine peaks, so it 'adjusts' to the cathode offset. But that, plus the voltage increase due to cathode current, allows more positive swing so the amplitude has to increase in order to get the 'average' bias level on the grid. At least, that's what I think is happening.

I forgot your previous numbers were peak to peak so I think I've still got too much amplitude. Is that right or is 6.6Vp small enough?

I first tried triode mode and got more current drain than I liked but that was aggravated by the high B+ so I should probably try that again to see if it's easier or harder to control.

I like the top end too but if you're using an AA5 cap there'd be enough range for the whole band. The problem is, if there's no slug you can't align it.

If I remember correctly that's about what I'm getting from my VFO, the 270pf coupling cap does drop approx a volt though... I used a 100pf though when the 12AU7 was in it with only about 3v pp... I'm not sure if the osc voltage is the reason or if it's a combination of other factors(probably Stereo Tool as much as anything), but this transmitter is sounding as good as anything on the AM dial(better than any of the WGOD stations)...



Sure you can, assuming the freq is high, just hang a 'lil more cap across it...

When I said shoot for the upper end, the plan is a "short" wind(similar to the 1V2 coil) and pull the freq down with larger cap value that would increase stability...

Tom,
Do you have a new schematic showing the changes you made? I am planning to sacrifice one of my old radios (I have about 10 or so) that will never get rebuilt for parts to build the transmitter.
Jerry W5JH

Okay, so 6.6Vp is at least 'in range', if maybe a bit toward the high side.

Well, uh, sure. But I know you know what I meant


Sure but you don't need a honker variable for that as you can just slap a fixed in to make up for the 'known' shortage. Whatcha gonna do? Pin the rotor to keep it out of the 160 meter band?

He he, when I first fired off the 1V2 coil it was up at 2350Khz with the tuning cap fully open(I was testing it on a Tx that has a AA5 tuning cap for freq adj), at mid point freq was down to 1500Khz... So yeah it's be a good Idea to add a parallel cap to prevent possibly operating out of band...

BTW any AA5 osc coil will be out of band with minimum std capacitance values, as the osc runs 455Khz "high" to compensate for the IF freq... In that light the "short" 1V2 coil wasn't that far off...



I haven't updated it yet, but will do so in the next few days... I'd sort of like to wind a easily duplicated coil and use that with a tuning cap... Any suggestions on a form that would be readily avail??? I've thought maybe a spindle from a CD stack would be easy for everyone... I have probably two pounds of #28 wire and some smaller as well...

Tom

Yeah, if you leave 'em tuned as in an AA5 but I think the 'universals' can be slugged on down. I'm not sure about that because I tie a fixed onto them and use the slug to tune a narrower range.

I forget, don't they usually use the 'small side' of the dual gang for osc?

On the 6CM8 pentode osc I upped my filter R to get B+ closer to target and at 165V, using the same values posted, the underside of the 1k is just a smidgen over 4V peak to peak (just under 1.5Vrms) with screen about 15V. Back at the 250V-270V range a 3.3 Meg screen resistor produced an even smaller output with 8.5V on the screen so it looks to me like that ought to work just fine. Although, I haven't checked to see what it does over the coil's tuning range because that AES thing doesn't have a 'screw' adjust and I can't find my hex slug tuner.

Even if we had to run it at a higher level, though, there shouldn't be a problem dividing it down, unless you see some sort of driving impedance into G1 issue.

Edit: Upon recheck, those numbers are with a 470 ohm under the cathode, not the original 1k.

Yep the small side, though the one in the Tx I was referring to has like sections and it's stamped 410pf... I'm guessing that's each section I've never measured it(didn't have a cap checker back when I built it), was from another Japanese radio I junked..




I know of one build that uses a padder cap for osc coupling to set the amount of drive into a 6888, I've never tried such but it seemed to work well enough...

Tom,
I wind coils on pill bottles. The pill bottles are readily available, usually free, and work pretty well. You can mount them using a screw through the lid or just put a screw through the bottom. You can drill holes in them for the wire to go through, etc. You might give one a try for the VFO coil. You could also use a type 2 toroid but a lot of people probably have not wound a toroid before and might be put off by it. I am not sure which type of coil would offer the best stability. What inductance are you looking for? I could wind one up and measure it for you.
Jerry W5JH

Well, as the saying goes, "since I had an oscillator," I went ahead and breadboarded up a 6GY6 transmitter with the 6CM8. Triode audio gain is arbitrary, about 5 (unbypassed Rk to reduce distortion), and B+, loaded, serendipitously came out 200V so I used the 2007 6GY6 values.

What I did discover is that too low an osc amplitude is worse than too much.

Our discussion had me a bit concerned about having too much so I cut it way down to 1Vpp (more like the components I had 'stuck' me there) and got positive mod flat tops long before even 50%. After diversions into Rk values it occurred to me, well of course. Them little electrons go through both sections and if there's not enough RF then positive audio can't 'let it through'. There just isn't any 'power' to 'pos mod' to. (One could, conceivably, slide bias way down on the G3 curve but that would increase non-linearity)

So I went back to the original pentode osc configuration which, because of the higher B+, gives about 6Vpp, I think it is, and that works fine.

I wouldn't say the components values are 'final' but the 6CM8/6GY6 pair is working great, even with just a wet noodle off the plate since I didn't have anything to do antenna matching with, so there's no reason why the others wouldn't work just as well.

I've done same, but for space considerations I'd like to use a smaller form, which is why I mentioned the CD case spindle... A screw through the center of the spindle could be used to hold it to the chassis or PCB... The coil I wound on the 1V2 tube is about as large in diameter as I want to use...

Maybe most would rather use a pill bottle, which is OK with me as I basically have my Tx built to my satisfaction and not looking to modify it... I'd just like to design a easy to build VFO circuit that won't cause a big headache for the less experienced builder...

As far as value I really don't know, as there seems to be a multitude of configurations in std BCB radios...



Hey that's great, basically the 6GY6 and similar 7 pin varieties are easy to work with and perform excellent in a small Tx application... A osc signal of around 4-6v PP seems to be the best for these types...

Tom

They do seem rather simple but you know "always got a question" me

But, for some background first, I was toying with the idea of a "resistance stabilized oscillator," although this isn't really the right configuration so I never got around to it. But it did get me to thinking about the screen. Screen current follows plate current so removing the bypass cap adds negative feedback. That clearly happens, since amplitude drops, so I lowered the screen R to 680k. Whether this does much, if any, actual 'stabilization' I don't really know but at least it saves a component.

So that had me checking mod again to make sure I hadn't messed something up. Well, seems to me the listed 470 ohm (with 200V B+) 6GY6 cathode resistor might be too low a value. But it could be I just don't understand how it should be adjusted, so I'm asking.

I'm just eyeballing the AM envelope but with 470 ohm I get pos mod flat topping somewhere around 80-90%. Maybe that's 'expected'? At any rate, that wasn't due to osc amplitude but audio trying to go positive G3 which, if we want idle centered on the whole G3 curve (maybe wrong assumption?), indicates bias is too low.

Frankly, by that point I had been futzing with it for a while and was getting tired of 'notch by notch' re-soldering things so I did a 'leap' change to see if it had a distinctly visible effect and went to 1k under the 6GY6. At that value I get what looks like a decent sine at virtually 100% mod. Although, it's a bit difficult eyeballing the '0' slope so 1k may be on the high side and maybe 820 would be better (680 ohm was still too low, though).

At any rate, the question is, for the notes on 'adjusting' that Rk for 'best linearity', just how is one to ascertain that?

He he, I been caught again... Actually I've been meaning to update Rk on those earlier schematics, but basically the 3v cathode in the notes applies... Usually 2.7-3.0v is best for Sylvania & RCA tubes while GE 6GY6 like closer to 2.3-2.5v... I listed 3v as erring on the high side doesn't effect modulation linearity to anywhere near the degree as being low... Rk value will vary mostly due to the screen current which also effects RF output... Using a resistance sub box, I've found best RF is usually with somewhere between 10 & 15K ohm on the screen or approx 2/3 of B+...

Oh yeah you were asking about Rk, I use a 1k pot to find the sweet spot that's usually 560-680 ohms, though one build I gave a friend needed closer to 800 ohms(but that one was using somewhere around 8-10v osc drive, so is running in at least class B)... The 470 was left over from when I was using a lower value screen resistor... To find the absolute best modulation without flat topping, I look at the detected audio while playing music, then tweak the pot... The varying amplitude makes clipping in either positive or negative direction obvious...

Tom

RCA, Sylvania, GE... You mean the label on the tube or the box?

Bwaaaaaaaaaaaahahahaha. Just kidding. The one you sent me is an RCA in a GE box

Anyhoo, that's what I mean. I'm seeing it needs more like 4.25V, maybe 4V if the (untried) 820 is ok. That's a pretty big difference. Could my gorgeous angelic sine, instead of that nasty old square wave, be affecting it?

That reminds me, I was rather surprised at screen volts. With the 10k I'm seeing 175V with B+ at 200V so if it's 'supposed' to be at 2/3's that might explain at least part, if not all, of the higher G3 bias.

What was that you said, Mr. Spock? "Very simple?"

Yeah, I need to find a simple way to get to the demod but my cell phone doesn't have AM, just FM.

Seems to me it maybe needs a 'two step' setup, like maybe first tweaking screen to some target and then Rk.

Btw, I added a variable to the osc tank and am seeing amplitude variation over the tuning range which, I think, is pretty much SOP but that aggravates the osc amplitude matter.

You have to watch out on late RCA tube they were generally mfg by GE as RCA stopped tube production in '76, after that the ones I've seen are all GE...

Well something is creating a difference, I honestly dunno where it lies but I've been using approx 3v on the cathode with great results when driving with a sine wave...

Bingo, a little tweaking on each goes a long way... This 6BK7-6DT6 Tx has a 10K supplying screen and now 680 for the cathode, k voltage is around 2.5v, seems the 6DT6A now in use likes more bias than the 6DT6 I was orignally using(Gm specs are different between these two tubes)...

Well, yeah, but the question remains, what is the screen target and, the ever popular, why? Like, is there a 'problem' with my screen at 175V with Rk 820-1k?