Hi,
I would like to try to follow You steps above as I have a power transformer to match a higher power RLD...
Some questions arise:
I have possibility for either 400V or 450V as anode voltage on the EL84. Any issues in chosing either?
What will be the optimal G2 voltage?
Normally it is seen as the same as Va, but here it will not be good to go beyond 300V. I am thinking halv value of Va. Looking in the datasheet is can be seen that there is a Ia dependency...but I guess this can be compensated by a lower G1 bias voltage (I am using fixed bias).
Generating G2 voltage:
1. Scale the +Va down with a series R-C that will drop the needed voltage. But this will make the G2 voltage depending of the IG2 draw and thereby the driving voltage....problem?
2. Making the last capacitor of the +Va supply a series connection of two 470uF/385V electrolytics. The top will then supply the anodes with 450V and the center connection the G2s with the half voltage: 225V...
Will this work?
Best Regards...
I think that your application is different enough from the original design intent that it warrants starting a new thread... but for some basic answers to get you started, I would recommend having an adjustable G1 supply per tube, and an adjustable G2 supply per channel at the very least, and even better would be per tube. In my experience you can put much more voltage on an EL84 plate than the datasheet recommends, but the G2 voltage is a bit more "particular" as I've had failures or runaway when you get higher than 375-400 volts G2-cathode... it also gets harder to set operating points reliably when you use G1 as the main adjustment, and it will take excessive negative voltage unless you go lower on the G2 voltage.For G2, an RC divider supply will not be adequate, some sort of regulator will be needed, but it needn't be terribly complex. A simple mosfet follower off of an adjustable reference would be plenty, or even a Maida regulator. For what it's worth I prefer using the 6V6GT or Russian equivalent 6P6S for higher volt operation, they seem more rugged than the 84 in this sort of application. Either way, you won't reap much benefit unless going to a higher load impedance to take advantage of the higher plate voltage, something like a 10-12k works well, I liked using an 8 ohm load with a 6000:4 transformer for a ~12k impedance when going with a 475 volt plate supply, and if I remember right my G2 supply was around 200 volts or so, adjustable from 150-330 volts.
(Perhaps a new thread for your build is in order, titled something like "EL84 higher voltage operation" or similar...)
(Perhaps a new thread for your build is in order, titled something like "EL84 higher voltage operation" or similar...)
Too easy to "red plate" an EL84 running it at the upper limits. I fixed an amp using EL84's where an output tube obviously red plated and collapsed the glass envelope. The 6AQ5 also was known for this.
Screen voltage supply can be regulated using a suitable dropping resistor off HT and inexpensive VR tubes... A 0A2 will regulate 150 volts; two OB2's in series will regulate 210 volts nicely. KISS principle applies.
6V6's or the rarely mentioned 6F6's appear to tolerate higher operating voltage better compared to the EL84.
Screen voltage supply can be regulated using a suitable dropping resistor off HT and inexpensive VR tubes... A 0A2 will regulate 150 volts; two OB2's in series will regulate 210 volts nicely. KISS principle applies.
6V6's or the rarely mentioned 6F6's appear to tolerate higher operating voltage better compared to the EL84.
Totally forgot about the VR tube approach, and it does work well, especially at the reduced screen voltages that would be used in this application. Not adjustable, but that can be worked with without too much fuss. A string of high stage Zeners word work for better voltage selection, or a mix of Zeners and gas tubes. It can easily be made blamelessly quiet with RCR filters, or a series current source, given adequate thermal management.
I think that the 6V6 (and 6F6, fantastic and underrated tube, IMO) will work better due to the larger plate structure being able to handle the heat and voltage better, the lower heater dissipation, and the further spacing of the glass to the internal structure. I happen to like the 6V6 a great deal, they seem more rugged than the EL84, even if they are lower GM and output, but they do sure sound fantastic when run at nice operating points, and the lower current heater is nice too. The main thing the EL84 has going for it is the low drive voltage requirements.
I try to rarely make such generalisations, but in my opinion (tetrode mode compared to pentode, at least) the 6V6 sounds cleaner and more comfy than the EL84 for a general use hi-fi amplifier. Most of my EL84 builds had more odd harmonic content, and a larger increase in higher harmonics once feedback was applied, and I think maybe that contributed to it, at least when run off the same supply and load, at operating points that put them on even footing.
I think that the 6V6 (and 6F6, fantastic and underrated tube, IMO) will work better due to the larger plate structure being able to handle the heat and voltage better, the lower heater dissipation, and the further spacing of the glass to the internal structure. I happen to like the 6V6 a great deal, they seem more rugged than the EL84, even if they are lower GM and output, but they do sure sound fantastic when run at nice operating points, and the lower current heater is nice too. The main thing the EL84 has going for it is the low drive voltage requirements.
I try to rarely make such generalisations, but in my opinion (tetrode mode compared to pentode, at least) the 6V6 sounds cleaner and more comfy than the EL84 for a general use hi-fi amplifier. Most of my EL84 builds had more odd harmonic content, and a larger increase in higher harmonics once feedback was applied, and I think maybe that contributed to it, at least when run off the same supply and load, at operating points that put them on even footing.
Never, ever, exceed the published dissipation limits! Within reason, liberties can be taken with the voltage specifications, but excess heat, AKA power dissipation, wrecks all things electronic.
As for "aggressive" designs, like Fisher's, my thinking is reliance was placed on tube builders shipping 7189s labeled as 6BQ5s. That was a fairly common occurence, given profitability considerations. These days, the only tube I would install in 1 of those "aggressive" designs is the Russian 6Π14Π-EB (6p14p-ev), AKA EL84M, which is a genuine 7189 equivalent. That variant combines the toughness of old rubber boots with pretty darned good sonics. New Sensor, you have a winner on your hands; please continue to produce the tube.
As for "aggressive" designs, like Fisher's, my thinking is reliance was placed on tube builders shipping 7189s labeled as 6BQ5s. That was a fairly common occurence, given profitability considerations. These days, the only tube I would install in 1 of those "aggressive" designs is the Russian 6Π14Π-EB (6p14p-ev), AKA EL84M, which is a genuine 7189 equivalent. That variant combines the toughness of old rubber boots with pretty darned good sonics. New Sensor, you have a winner on your hands; please continue to produce the tube.
I have a variant of my SPP board (modified for fixed bias) that runs "EL84's" from 425 volts and makes 25 to 30 watts per pair into 6600 ohms. It's been running off and on for about 5 years on the same set of tubes. Those tubes were removed from a Baldwin tube organ that was pulled from church use....."only used on Sunday morning." They are actually Baldwin branded Sylvania 6BQ5's from the 60's.
The key to this "abuse" is to keep the screen grid voltage below 325 volts and keep the idle dissipation down to 9 watts or so, roughly 20 mA. If the plate dissipation hits 15 watts or so on peaks, this is OK for music since those peaks are short and infrequent. This may not be the best choice for an amp that will see continuous sine wave testing, but we don't usually listen to continuous tones for long. I play guitar through mine on occasion and push it well into clipping.
I made a simple zener diode / mosfet "regulator" for the screen grid that is adjustable from 200 to 350 volts and tweaked it to maximize plate efficiency at about 10 to 15 watts of power output. I don't know the exact voltage since the amp hasn't been opened in years.....it just works.
When I was building this thing I noticed that most old 6BQ5's worked fine, as did JJ EL84's. Some of the Russian and Ebay sourced tubes red plated. I wound up usinc the Baldwin branded Sylvanias. This was done 5 or 6 years ago, so the new production tubes available today could be different.
The key to this "abuse" is to keep the screen grid voltage below 325 volts and keep the idle dissipation down to 9 watts or so, roughly 20 mA. If the plate dissipation hits 15 watts or so on peaks, this is OK for music since those peaks are short and infrequent. This may not be the best choice for an amp that will see continuous sine wave testing, but we don't usually listen to continuous tones for long. I play guitar through mine on occasion and push it well into clipping.
I made a simple zener diode / mosfet "regulator" for the screen grid that is adjustable from 200 to 350 volts and tweaked it to maximize plate efficiency at about 10 to 15 watts of power output. I don't know the exact voltage since the amp hasn't been opened in years.....it just works.
When I was building this thing I noticed that most old 6BQ5's worked fine, as did JJ EL84's. Some of the Russian and Ebay sourced tubes red plated. I wound up usinc the Baldwin branded Sylvanias. This was done 5 or 6 years ago, so the new production tubes available today could be different.
Yup. High anode voltage, moderate G2 voltage, and less negative G1 voltage with a smidge higher plate load impedance works great, is more linear/lower THD, and is easier to drive than high plate voltage, high G2 voltage, and more negative G1 voltage.
On pentode/tetrode types G2 is much more effective than G1 at controlling plate current, so using this to your advantage will allow you to run G1 less negative than you would had you set G2 higher and closer to the plate voltage.
On pentode/tetrode types G2 is much more effective than G1 at controlling plate current, so using this to your advantage will allow you to run G1 less negative than you would had you set G2 higher and closer to the plate voltage.
Having adjustment for G2 per output pair is usually good enough, if going for individual G1 adjustment or balance to fine tune things. On most applications G2 current will not be all that high, so the expense of setting up and heatsinking an additional regulator isn't all that much fuss. How gilded do you like your lilies?
Thanks for this thread, as it's germane to the EL84 amp I've yet to finish. Its power transformer came from a Fisher console, and the B+ is a bit frisky, and may remain so under load with having to supply just four EL84s (the fnont end is sand-state and runs off a separate supply XFMR). I'll probably try using a zener-regulated screen supply for each channel, or possibly a simple fet-based linear pass solution. I'm using JJ EL84s.
It is not about EL84 only. It works with any pentodes and beam tetrodes. For example, I have a PP amp with paralleled 6P3S tubes. 40W only from 4 tubes, but it was the only way to get high-end sound from such tubes. When G2 voltage is as high as anode voltage, for such linearity they would dissipate enough power to melt down the glass.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.