• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Plate curve calculations - am I doing this right?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

I'm looking at the 2A3 plate curves. My operating point is 290V plate-cathode voltage, 62mA plate current, about 55V bias voltage (cathode biased). I plotted this point on the plate curves (and drew the 2500 ohm load line), and am trying to calculate u, gm and Rp. Rp is the slope of the plate curve at the operating point, so I drew a line that's tangential to the nearest curve at the operating point. That gives me a value of around 660 ohms, which is close enough to the 800 ohm nominal value found in the data sheet (which is probably valid at a different operating point). I've read advice that it's more accurate to calculate u and then get gm from u and Rp, than trying to calculate gm straight from the plate curves, because the lines are further away in the vertical direction. So, to get the value of u, I draw a horizontal line through the operating point, intersecting the nearest plate curves on either side. This gives me a plate voltage change of about 40V, for a grid voltage change of 10V, or u = 4, and so gm = u * Rp = 2640.

Do those numbers look right? Does a 2A3 at this operating point have a voltage gain of just 4?

The hum balance pots on my amp have taken care of almost all the 60Hz hum, I have some 120Hz hum still left though, and I'm thinking of trying the Tubecad idea of adding a capacitor from B+ to cathode to act as a voltage divider with the cathode bypass cap to inject the right amount of B+ hum into the cathode so that it cancels out the hum on the plate (article link: http://www.tubecad.com/april99/page2.html). To do this, I need to know the value of u.

On a related note, is this circuit pretty similar to the ultrapath connection circuits I've seen? I've seen some that remove the cathode bypass cap and only leave this cap from cathode to B+. Those talk about making the output circuit loop smaller and not making the signal go through the last PS cap. They don't talk about cancelling hum, and the appropriate cap value is usually describes as "from 7uF to 100uF, about 40uF works best", at least in the few posts I've read. I'm wondering if both circuits do similar things?

Thanks in advance,
Saurav
 
Hi Saurav,

Your reading of the plate curve data is correct - sounds like you have that nailed! If you had called up 2A3 data you would see mu quoted as 4.2 at normal operating points and as you can see by the curves your slightly hotter than traditional point is close to that.

On the other question, Ultrapath and the Tubecad hum reduction technique are not the same and don't do the same job. As you state Ultrapath bypasses the powersupply and cathode bypass cap to create a short and well defined ac signal return path. As such it doesn't do much to reduce hum. The tubecad technique does directly affect hum by introducing a measured amount (via the cap divider) of anti-phase hum into the valves amplifing path and so providing active cancellation - to a degree. It is possible to get better than 20dB of cancellation by this 'trick'. Funny how a capacitor in the same place can do two jobs according to it's value... (It does, of course do both jobs but it does one better than the other according to how you choose it's value)

Hope this helps

ciao

James
 

PRR

Member
Joined 2003
Paid Member
> I know that preamp/driver tubes will have a higher mu, but I didn't know that power tubes were this low.

For a given heater power, a tube's Gm near zero-bias is pretty much "fixed". Some of the 1960 tuner tubes do a little better than the ancient 1930s tubes, but not a lot.

The plate resistance is Mu/Gm. We should mind the difference between large-signal Rp and the incremental Rp, but for most triodes the difference is not large and the trends are valid.

For best power from a fixed supply voltage, we want the lowest possible Rp. But for a given heater power, Gm is roughly fixed. (Assuming your grid design is not stupid, Gm is proportional to cathode area, and heater power is proportional to cathode area.) So the only way to get a small Rp is to move the grid and plate to get a low Mu.

You use Mu=20 to 100 for voltage gain where current and power output is small. But for POWER, you must use low Mu tubes and accept the low gain (high drive voltage).

But Mu=1 or Mu=2 is absurd. The required drive voltage is larger than the plate supply on the output stage. Since we usually design the power supply for the output stage, since it takes the most power, it is annoying and costly to add another higher-voltage power supply just for the driver.

If you assume some rules-of-thumbs: peak RC-coupled driver output is 0.20*B+, output stage bias is about 0.63*(B+/Mu), then the lowest practicable Mu for a tube with R-C coupled driver on same supply as the output is about Mu=3.

You hate to whack a driver that hard, and you would like a cheap R-C ripple filter on the driver (so the main B+ does not have to be as clean as the driver B+). Then the optimum is about Mu=5.

All the classic audio power triodes have Mu=4 or 5, in that range: 2A3, 300B, type 50, etc.

(Pass Regulators work a little different and can use a lower Mu, such as 6080/6AS7 Mu=2. The difference is that the pass-tube only sees a small part of the total power supply voltage, but the reference amplifier can use the whole supply voltage.)

In a somewhat related field, TV Vertical Sweep amplifiers, the power triode on a wide-deflection design runs a Mu of 6 or 7. See 6EM7. They need big current and high efficiency at low suply voltage, but need more-than-minimum gain to keep the sweep feedback working at extremes of voltage and current. (There is an older class of V-Sweep tubes with Mu=20, but these were for narrow-angle CRTs that did not need huge yoke current, or for TV sets that had large supply voltage available.)

And interestingly: the huge TV Horizontal Sweep pentodes also have Mu=5 or so when triode strapped. They needed maximum efficiency, so they needed low large-signal plate impedance, which in a pentode is approximately the triode-strap plate resistance.

OTOH, the venerable 6L6 and its many kin run triode-Mu around 10 or 20. Maximum current is a little less, but the pentode mode allows this current to be available at fairly low plate voltage swing, and the higher Mu means less driver voltage and distortion.
 
diyAudio Senior Member
Joined 2002
Hi,

Interesting post but...

But Mu=1 or Mu=2 is absurd. The required drive voltage is larger than the plate supply on the output stage. Since we usually design the power supply for the output stage, since it takes the most power, it is annoying and costly to add another higher-voltage power supply just for the driver.

Frankly, I don't much care about that since I feel it to be a better approach to use independent supplies, some even regulated, for input and outputstages.

From your reasoning all CF outputstages are out since they don't have any gain at all, in fact their gain is even a negative figure.
Or are you just reffering to the stated mu in the datasheets which is a topology dependent figure anyway?

Anyway, I feel you are always better off keeping PSUs for different stages as separate as possible and I always use totally independent supplies on a pro stage basis...obviously I don't care about the cost, just the performance.

Cheers,;)
 
Ex-Moderator
Joined 2003
Pedants unite!

Tim, mind you, in dBs, CF does have negative gain (even more pedantic).

PRR, your points about output valve mu were very interesting, and I'd not thought about it in that way. It's always nice to know why things are the way they are. Thanks for that.
 

PRR

Member
Joined 2003
Paid Member
> Interesting post but... Frankly, I don't much care about that since I feel it to be a better approach to use independent supplies, some even regulated, for input and outputstages.

I am just spouting "Conventional Wisdom". You are of course free to be as wild and crazy and extravagant as you want.

But what do you gain by ignoring my advice? If you go to a very low-Mu tube you get a little more B+ current and output power for a given supply power. OTOH if you pick a high-Mu tube, it drives a little easier but needs higher B+ voltage and/or more heater power for a given output power. But your philosophy does not care about power supply costs. I'm sure you will happily build whatever size PS a stage needs.

And nobody is going to build a special-Mu tube just for you. The mainstream of hollow-state electronics is and always has been "one B+ supply", direct to the power-stage and decoupled to the driver and earlier stages. In that crassly commercial world, Mu= "about 4 or 5" is the only way to go for profitable power. With rare exceptions (pass-tubes the main class) you won't find many high-power tubes with Mu much different from 4-5 for triodes, 5-20 for tetrodes. Even the advent of "unprofitable power" ($5,000 for 2 watts???) has not stirred tube designers to cook-up any new tubes. (Indeed they have a hard time just imitating the old tubes.)

Anyway, if you go too far from the time-proven parameters, you do get insane fairly quick. If you are willing to build multiple supplies, you are not the kind of designer to accept rules of thumb blindly. When I say a driver voltage amp can make peak voltage about 20% of its supply voltage, that's at 5% THD or so, which is rather high. An extravagant design would aim for lower THD, which means the peak audio voltage will be a smaller percent of supply voltage. Peak of 2% of B+ gives about 0.2%-0.5%THD.

So say you pick a power tube with Mu=2 (for low plate impedance) and B+ of 250V (6080/6AS7). Grid bias and peak grid drive will be about 80V to 100V (more like 100V for 6AS7 to keep it from melting at 250V). If the driver can only swing 2% of its supply for low THD, the driver needs to run on 5,000V!!! Even if you don't care about the cost of a 5,00V supply, that will limit your choice of tubes and you may not find one you like with that kind of voltage rating.
 
Ah yes, forgot about dB. :clown:

Hmm not sure why but I feel like putting on my psychologist hat:

I'm sure we could go on for weeks about this. That it could seems to indicate either a. interesting and constructive discussion taking place, or b. someone's hit a nerve with someone else's beliefs. Since diyAudio is often more belief (12AU7, need I say more) than engineering (when's the last time someone here gave a darn about .0001% THD), b. is easily possible.

So put it to rest. (As much as I might like to jump in the discussion.)

Tim

P.S. Psst... pentodes have lower distortion than triodes... :devily:
 
...Ouch! The truth hurts...

P.S. Psst... pentodes have lower distortion than triodes...

True - sometimes...

But we don't listen to THD very well but we do listen to the distortion spectrum exceptionally well - that's why a lot of people like listening to DHT SE.

but again it is perfectly possible to engineer a pentode amplifer to have low distortion and a nice distortion spectrum... but it's easier to get a nice sounding DHT SE and it's fashionable.

PBR.

Interesting posts - Thankyou. And you reinforced why I don't like resistor loaded drivers :nod:

A question - In terms of higher mu triodes being easier to drive ...how do you balance miller capacitance against gain ?

ciao

James
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.