Apparently. You see this in the oldest amps. As magnets got bigger, and apparently as HumBucker pickups got popular, most g-amp inputs switched to cathode bias.
OTOH Bogen PA mike inputs were gridleak bias to the very end. Signal levels (before rock bands and before $99 hi-output condensers) were reliably 10mV or less. (Contrast to 50mV rising over 200mV for e-guitar.)
A well-picked gridleak stage will tend to find the most gain for the least pennies, including lifetime costs (e-caps go bad).
All the HIGH-gain small triodes were brewed to reliably throw the grid negative for Rg in the 1Meg-22Meg range. This could fail if tube current got huge-- 100K and up (to 470K) plate resistor ensures that won't happen.
The "big" trouble is you won't get huge output signals. The op-point must be low current. The plate idle voltage is known in general but not firmly fixed. You will find plates sitting at 50V or 150V. If you only need a 7V peak output (as to drive a table radio's final), this makes little difference. If you need 60V peak to flog a 6550, both the uncertain op-point and the low current doom your design to trouble.
Yes, the 2nd stage of a NFB phono preamp is a groovy place for gridleak. Output level can't be all that high relative to available supply voltages. (In part because a 2-triode RIAA can't make large gain.) If you need 5V peak output (on a HOT record), and V2 runs gain of 50, that's only 100mV at V2 grid, which is within the zone where gridleak works fine. The high grid resistor makes the most of V1's gain, which is always precious to hit the 50Hz pole.
I started my long fascination with tube amps in the mid 60's with a hand traced schematic of a Fender Champ 5C1. It used a grid leak biased 6SJ7 driving a 6V6GT powered by a 5Y3. I built these with parts fetched from the trash dump, but 6V6 types were not too common. 6K6's, 6W6's, 6Y6's and 6BQ6GT sweep tubes were more common and I tried them all. I never understood why the bigger 6DQ6 wasn't louder, but I didn't yet understand things like bias, impedance and load lines. 6SK7's work as the input tube with different sound and overload characteristics.
The grid leak resistor isn't too critical and most everything I tried from 1 meg to 10 meg worked by adjusting the screen resistor value to get about 100 volts on the plate. Note that you can't just stick your voltmeter probe on the grid since it's input resistance is often lower than the grid leak resistor. I had an Eico VTVM built from a kit that featured an 11 meg input resistance, but even it would change the bias.
Grid leak bias falls apart when the stage is driven close to, or into clipping. Plug your germanium fuzz box into that 5C1, crank it up full, and you can drive that 6SJ7 into grid blocking territory. The long time constant afforded by megohm valued grid resistors can leave the amp silent or severely distorted for a second or longer.
The optimum grid leak value does depend a lot on tube to tube variation and condition of the tube. Tubes with a wee bit of gas that would operate fine in a cathode biased design may exhibit positive grid voltage and silence with a 4.7 meg "leaker."
Want to use a common cathode tube, don't like electrolytics? LED's are your friend. Ordinarily an LED needs several mA to get somewhere near a constant voltage drop, so add some "boost" current with a resistor to B+. A red LED can be found that lights up on 1.7 to 2.2 volts and this will work with several tube types including the 6SC7.
The grid leak resistor isn't too critical and most everything I tried from 1 meg to 10 meg worked by adjusting the screen resistor value to get about 100 volts on the plate. Note that you can't just stick your voltmeter probe on the grid since it's input resistance is often lower than the grid leak resistor. I had an Eico VTVM built from a kit that featured an 11 meg input resistance, but even it would change the bias.
Grid leak bias falls apart when the stage is driven close to, or into clipping. Plug your germanium fuzz box into that 5C1, crank it up full, and you can drive that 6SJ7 into grid blocking territory. The long time constant afforded by megohm valued grid resistors can leave the amp silent or severely distorted for a second or longer.
The optimum grid leak value does depend a lot on tube to tube variation and condition of the tube. Tubes with a wee bit of gas that would operate fine in a cathode biased design may exhibit positive grid voltage and silence with a 4.7 meg "leaker."
Want to use a common cathode tube, don't like electrolytics? LED's are your friend. Ordinarily an LED needs several mA to get somewhere near a constant voltage drop, so add some "boost" current with a resistor to B+. A red LED can be found that lights up on 1.7 to 2.2 volts and this will work with several tube types including the 6SC7.
;-)
Love it. Whetstone bridges — tho' a bit of a pain to set up — are the answer. The idea being that one can produce an essentially near-infinite impedance measurement of a DC voltage in a WB configuration. Have to fiddle with the 10 turn wire wound bridge balance tho'. AS the measurement(s) come to zero balance. With practice, a good measurement can be had in less than 15 seconds. Might take a minute or two tho' the first time you try!
Ah, the smell of Metrology so early in the day…
GoatGuy
Tubelab_com and GoatGuy,
Oh, for those old fundamental measurement days!
I regret never owning a VTVM. I went from a bare bones 1mA meter movement (roll your own measurements), to a VOM, to a DMM.
Fluke made a living off of a box that 'essentially' had nothing but a calibrated DC power supply, multi-decade resistor voltage divider, current limiting resistor, and a somewhat robust zero center galvanometer (treat with care, or there was more Bust-ed than Row-ing). The Navy and other military had scores and scores of these. Infinite impedance here we come!
I had customers call me who wanted to set one oscillator frequency exactly the same as another oscillator frequency, but they did not have a frequency counter. They did have a 2 channel scope with XY. Problem solved!
And other countless measurement stories, especially from me being in the middle of the Pacific Ocean.
Oh, for those old fundamental measurement days!
I regret never owning a VTVM. I went from a bare bones 1mA meter movement (roll your own measurements), to a VOM, to a DMM.
Fluke made a living off of a box that 'essentially' had nothing but a calibrated DC power supply, multi-decade resistor voltage divider, current limiting resistor, and a somewhat robust zero center galvanometer (treat with care, or there was more Bust-ed than Row-ing). The Navy and other military had scores and scores of these. Infinite impedance here we come!
I had customers call me who wanted to set one oscillator frequency exactly the same as another oscillator frequency, but they did not have a frequency counter. They did have a 2 channel scope with XY. Problem solved!
And other countless measurement stories, especially from me being in the middle of the Pacific Ocean.
Thing about them, is that they were so robust. You could accidentally spark 250 kV on their inputs, and while you might blow a fuse, the input circuit would take it without so much as a wee wisp of smoke.
Compare that to a solid-state input meter. I know of none that wouldn't get their skivvies in a twist over a 100-plus kilovolt input accident.
You just had to keep around though a smallish box of really nice gold-plated terminal input tubes for the aging issues of tube circuits. That an a few “references” of different sorts to field calibrate the units. Again, field-calibration only took a couple of minutes once you get the hang of it. And if you “did it yesterday”, then without any fanfare at all, you could depend on the VTVM meter readings after warm-up/power-up.
Interestingly, the circuits used also didn't need extensive warm-up. The built-in jeweled suspension meters (with anti-parallax strip mirror and ultra-thin needles) could easily deliver ±1% readings across 50% of the scale. And the VTVM (at least my HP) would “settle down” in about 1 minute of warm-up. I know — I tested hundreds of well calibrated resistors and voltage sources with 'em. 1% was easy.
And mostly anything better than 2% was mostly overkill. 5% was the norm or MILSPEC circuits of any delicacy. 1% for RADAR ranging, 10% or 20% for all the electrical-engineering power-and-sigops bull.
The near-total immunity to input overload was also the much repeated reason why both the US Army and Navy resisted going “solid state” for so long: EMP (electromagnetic pulse) nuclear war issues were sidestepped with high confidence.
GoatGuy
Every one I ever fixed had a bog standard 12AX7 and a 6AL5. Some of the really old ones used a 6X4 rectifier.
I really didn't understand too much in my pre-teens and early teenage years, but that didn't stop me from building stuff, blowing it up and fixing it. I convinced my parents to skip the usual Christmas presents one year and get me that Eico VTVM kit. It had to be around 1960. I built it and used it for a long time.
I started working at Motorola in 1973 on the HT220 radio test and tune line. Every equipment rack had an RCA Senior Voltohmyst VTVM. In 1975 I moved to the calibration lab where I got to fix test equipment. The VTVM's lasted into the early 80's, eventually being replaced with Fluke 8000's. That's when the weird stuff started. The old VTVM's were relatively immune to RF. We are talking about a place that made police walkie talikie. Everybody above the assembly line workers carried one including me. Every radio shipped was tested with a live antenna at least 4 times before it went into the box. That's if it passed every test on its trip through the factory, which almost none did. Key up a 5 watt VHF radio near a digital meter, random digits happen.
Those things lived forever in the factory since every station was fitted with 30 dB pads. Engineering was a different story. Rookie engineers would zap those things so often that HP would sell us the tiny thermistors so we could fix them ourselves. At first we used conductive epoxy, then I figured out how to fit the bolometer base into a microelectronics wirebonder and weld the little wires in place. It's when the rookie engineer keys a 5 watt radio directly into a brand new HP system 7000 spectrum analyzer that the bosses want to fire someone.
I had a HP audio signal generator with a picture tube and 1 KHz camertone generator for calibration. Another similar generator that I disassembled for parts had a salt crystal in a vacuum tube instead of a camertone.
All my VTVMs had 12AU7.
Since most used 200uA movements, a 12AX7 "could" pull the load; but 12AU7 pulled it easy.
Yes, one way to read the grid voltage without loading is to contrive an adjustable low-volt source (can be a C-cell and a 10K pot), run voltmeter from wiper to pot, adjust pot for null on meter. Now disconnect and read the wiper voltage. Wheatstone, simpler.
However if you do this a lot, just rig a TL072 voltage buffer, no bias resistor (but put 10K in series as protection). The '07x's input current is much-much less than any tube's grid current until you get to Electrometer and Fancy German Microphone tubes.
Since most used 200uA movements, a 12AX7 "could" pull the load; but 12AU7 pulled it easy.
Yes, one way to read the grid voltage without loading is to contrive an adjustable low-volt source (can be a C-cell and a 10K pot), run voltmeter from wiper to pot, adjust pot for null on meter. Now disconnect and read the wiper voltage. Wheatstone, simpler.
However if you do this a lot, just rig a TL072 voltage buffer, no bias resistor (but put 10K in series as protection). The '07x's input current is much-much less than any tube's grid current until you get to Electrometer and Fancy German Microphone tubes.
Paraphrased, but I like it.
I vexingly often forget just how high the impedance is of “modern” JFET and similar not-marketed-as-such op-amps is. 10¹² to 10¹³ Ω. Wow! And for 55¢/qty1 at Mouser. Check out
for instance.
Nominal ambient temperature input bias of 1 pA; 25 pA at 85° C. 10¹³ Ω (that's 10,000,000 MΩ!) input impedance at 25° C. Just amazing. Electrostatic input protection ain't great, so probably better to just go with the JFET Texas Instruments TL072 parts. More robust.
Still… the things available today… are amazing.
GoatGuy
I've been using the leaded versions (not SMD) for cathode bypass caps in my DIY valve guitar amps for a couple of years now. They work fine, just as same value electrolytics would, but one hopes the multilayer ceramics will be more stable over the years.
I have used up to 22uF multilayer ceramics for high gm valves that require a correspondingly higher value of bypass cap to maintain the same low frequency cutoff.
The 22uF leaded versions are still not much bigger than the head of a matchstick, if anyone still remembers what a matchstick is. Was. They seem nearly extinct now.
-Gnobuddy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.