Parallel triodes

[mauiboy81]

Hey all, it’s been a minute since I’ve been building but have been diving back in recently. I have a question about parallel triodes specifically on the input stage…

The scenario is running parallel plates and cathodes but using separate inputs so you can either run into one or both of the triodes simultaneously. When only using the input of one triode and the grid of the other is grounded are the Rp and Rk that the triode “sees” double the actual values of the resistors like when running both triodes in parallel or does the single triode “see” the actual value of the resistor?

So if using 110k Rp and 1.5k Rk with parallel triodes it’s “seeing” 220k Rp and 3k Rk. If plugging into only one input is it still “seeing” 220k Rp and 3k Rk or is it the actual resistance of 110k Rp and 1.5k Rk because one triode is not working?

Hope that was clear. Thanks for your input!

[sluckey]

You have the basics backwards. You need to half the Rp and Rk values when paralleling triodes.

[martin manning]

The triode with the grounded grid won't be inactive. Its Vg and Vk are fixed, but it's plate current will vary with the fluctuating plate voltage, and it will be out of phase with the other one, which will cancel some of the output signal.

[mauiboy81]

Do I have it backwards? Using 110k Rp and 1.5k Rk with parallel triodes is like using 220k Rp and 3k Rk on a single triode. This is correct, right?

[mauiboy81]

Thanks guys. Martin, will the output signal cancelation vary at different frequencies depending on the voltage fluctuation or will it just be lower output across all frequencies?

[cdemike]

Do I have it backwards? Using 110k Rp and 1.5k Rk with parallel triodes is like using 220k Rp and 3k Rk on a single triode. This is correct, right?

You're correct, so long as the triodes are actually operating in parallel. As Martin pointed out, if you're just feeding signal to one, you'll actually have less signal than if they weren't sharing a plate resistor. For example, a triode gain stage with a 100k rp and 1.5k rk will produce more gain than an parallel triode setup with the equivalent 50k and 750R values if the parallel setup has only one grid receiving signal. So yes, the equivalents there are correct as long as both grids are receiving signal. Unfortunately, that also means this isn't the case:

So if using 110k Rp and 1.5k Rk with parallel triodes it’s “seeing” 220k Rp and 3k Rk. If plugging into only one input is it still “seeing” 220k Rp and 3k Rk or is it the actual resistance of 110k Rp and 1.5k Rk because one triode is not working?

What's the design goal with the parallel stage? If you're aiming for a high/low sensitivity input setup, switched cathode bypass caps may be more practical if you're looking to do something beyond the usual voltage divider and grid stopper approach. I did that in a recent build that also involved a parallel triode input stage, and I'm pretty happy with it (mine was for lower noise rather than higher gain). What I ended up doing was putting a resistor in series with a cathode bypass capacitor and having a Cliff-style switched jack bypass the resistor direct to ground on the sleeve switch. Of course, that'd also work with a vanilla single-triode gain stage, but you'd lose the benefits like lower noise or more gain potential.

[martin manning]

will the output signal cancelation vary at different frequencies depending on the voltage fluctuation or will it just be lower output across all frequencies?

Reduced output across all frequencies.

[sluckey]

Do I have it backwards? Using 110k Rp and 1.5k Rk with parallel triodes is like using 220k Rp and 3k Rk on a single triode. This is correct, right?

That's correct, but that's not what you said in the original post.

[mauiboy81]

Cdemike, the goal is high/low sensitivity inputs with lower noise on the high input and possibly a different flavor. And I was wondering if when running the low input through the single triode if that triode would be operating as if it had the 110k plate and 1.5k cathode or if it would still be like using a 220k plate 3k cathode because it’s paralleled with the triode that’s not getting an input signal.

[mauiboy81]

Sluckey, I thought that is what I said in my original post. Maybe not. I find it difficult to get these things out of my brain and write them down so they make sense. lol

I appreciate your input.

[mauiboy81]

Cdemike, Regarding the switched cathode caps I have done that on other stages with great results. I tried it once on an input stage of a silverface bassman and it hummed like crazy with no cap. I have read that it’s best practice to fully bypass an input stage for that reason but I also know that it is done. I might have to try it again.

Edit: on the silverface bassman I rewired the tonestack to more typical configuration and used the deep switch for the cathode cap. I ended up bypassing the first stage and then using the switch for the second stage bypass cap and like it a lot.

[cdemike]

Interesting that happened with the Bassman. Hiwatts are known to be particularly quiet amps, and most schematics for Hiwatt 4-hole preamps show an unbypassed 220k/2.2k input stage on the normal channel. My recent build is also the quietest Marshall-style amp I've played with that bypass cap setup, and I also borrowed the "gain control" from Sluckey's "Smoky" amp in a Hiwatt build which is also pretty quiet with the first stage essentially unbypassed; I suspect there might have been something else at play. With an old amp like that silver Bassman in a humid climate like Maui, I almost wonder if there may have been an issue involving oxidation on the old factory ground contact points (I said, having never laid eyes on the amp -- my wife would be roasting me right now...).

In the setup you described, the low input would have less gain than if it were set up as a single triode setup with identical rk and rp values, as Martin pointed out. It could still be a useful sound, and it's basically operating as a common-plate mixer in that configuration with one side just not having a signal to be mixed. Still I'd be concerned about the low output from that side. It could be worth checking into dissimilar triodes like an ECC823 or ECC832 if you're set on using two triodes as your input stage. Or, if the lower noise is a key piece, a more conventional split triode setup using a 12BZ7 would give you performance nearly identical to a parallel 12AX7 while still allowing you to isolate each channel and set them up differently. I got a handful of 12BZ7s a few months ago for less than $10 each, and I've had a ton of fun experimenting with them. As a word of caution, many 12BZ7s won't work well for input stages due to microphony, so this may be moot. I haven't tried the workarounds people try for other microphonic tubes like EF86s like rubber dampers on the tubes, but it seems like that could tame some of them.

[bepone]

why not 2 ecc83 halves with separated Ra, Rg, Rk.. connection joint after each anode coupling capacitors...

applications are many, dissimilar Rk (one high, one low), cathode bypass caps, using only one iput grid or two.. signal is only one or summed from 2 inputs..

[sluckey]

Sluckey, I thought that is what I said in my original post. Maybe not. I find it difficult to get these things out of my brain and write them down so they make sense.

It's easy to know what you are thinking but twist it around when you write it out. OTOH, it's easy to know what you are thinking but twist it around when you are reading. I'm guilty of both!

[mauiboy81]

It's easy to know what you are thinking but twist it around when you write it out. OTOH, it's easy to know what you are thinking but twist it around when you are reading. I'm guilty of both!

I will say I’m envious of how you guys can communicate technical concepts so clearly and concisely.