Hi all. I have collected a variety of parts that I intend to use in a KT120 single ended triode amplifier with a switch for ultra-linear operation. I have a pair of Edcor CXSE25-5K-8 Ohm Output transformers, and several power transformers that I can use. I have two Freed power transformers rated at 540V-0V-540V and 250ma, a single power transformer from an organ with 380v-0v-380v and 250ma, and a pair of NOS power transformers from the late 60s with 300v-0v-300v and 120ma. The Freed doesn't have filament windings, but I do have a separate Freed filament transformer. I also have several chokes that are 10H and able to handle 250ma each. Also, I have a variety of capacitors that will be able to handle whatever voltage I choose to use. I have been thinking about building the Alex Gendrano KT88 amplifier and upping the voltage and current to better suit the KT120, but any alternate design ideas would be welcomed as long as they are single ended triode/ultra-linear. I am hoping to get as much wattage as possible while having great SQ and while using only one output tube per channel.
I am hoping to get some advice from you all with respect to what you would do given the selection of components I have available. Freed with a choke input? Another one of the transformers with a capacitor input? Freed with a capacitor input for very high voltage? also, why would you choose the configuration that you suggest? There are many options, and a huge number of voltage and current combinations, as well as configurations for the power supply.
I already have an amp based on Alex Gendrano's KT88 SET design that used most of the power supply components from a organ chassis, and Edcor 15 watt OPTS. This sounds great and will take KT88, EL34, 6L6, 6550 etc. I want to build this one for KT120 specifically, with a switch that will allow me to also use KT150 when I get around to ordering them.
Any advice and opinions would be much appreciated.
I am hoping to get some advice from you all with respect to what you would do given the selection of components I have available. Freed with a choke input? Another one of the transformers with a capacitor input? Freed with a capacitor input for very high voltage? also, why would you choose the configuration that you suggest? There are many options, and a huge number of voltage and current combinations, as well as configurations for the power supply.
I already have an amp based on Alex Gendrano's KT88 SET design that used most of the power supply components from a organ chassis, and Edcor 15 watt OPTS. This sounds great and will take KT88, EL34, 6L6, 6550 etc. I want to build this one for KT120 specifically, with a switch that will allow me to also use KT150 when I get around to ordering them.
Any advice and opinions would be much appreciated.
Just wondering if anyone is running the KT120 and/or KT150 tubes at close to their dissipation limits of 60 and 70 watts respectively, as I would really like to get as many watts as possible out these tubes in a SET amp, with ultralinear as a switchable option. Another member has commented that he has experienced the start of red plating, on parts of the KT150 plate, at 70% of their 70 watt dissipation limit (50 watts).
Please check out this thread: http://www.diyaudio.com/forums/tubes-valves/248486-comparison-kt150-versus-kt120.html
es345 provided quite a bit of measurement data for SE near the end of the thread. In any case, it's never a good idea to run the tubes at their maximum rated dissipation limits unless you have a big budget and do not mind replacing tubes on a regular basis.
es345 provided quite a bit of measurement data for SE near the end of the thread. In any case, it's never a good idea to run the tubes at their maximum rated dissipation limits unless you have a big budget and do not mind replacing tubes on a regular basis.
Given 5K 60W 25W...
547VDC 110mA is 60W dissipation, as Pentode (and maybe in UL) should give ~~24W Sine audio in 5K load.
Add 10%-20% for filtering and self-bias. (Never run a valuable bottle near 99% Pdiss in fix-bias.) 600V to 650V. (Costly cap array.)
Add 10% for G2 and driver current, 121mA. Double for stereo, 242mA.
The 120mA PT is right out.
"540V-0V-540V" gives 750VDC cap-input, 480V choke-input.
"380v-0v-380v" gives 530VDC cap-input, 340V choke-input.
None of these options are real good fits.
As a hack, 750VDC dropped to 650VDC through a 100V 242mA = 400 Ohm 24.2W (use 50W) resistor. This can give great B+ filtering with modest caps and no choke (sorry!). The added 24W of heat may not be a big deal when you have 120W of plate heat and 32W of heater-heat (total 152W) already on board. But it will need a LOT of air-space around that resistor.
This suggests hollow-state rectifier. 100V of voltage waste is not uncommon for these. Hmmmm.... 5U4 and 5AR4 are a wee bit over their ratings. PY500 is not showing any red-flag, but I never heard of it. 6AU4 also seems to work (630V out) and that one used to be super-common. Good, because it is one-diode so you need two (and a 6V 4A well-insulated winding). And I am not sure PSUD is telling me a truth. There are a billion big diodes, but if you get one which will take the surge current it may not have the drop. Of course: we are asking the rect to be a 24W resistor, and few ordinary bottles can throw 24W.
547VDC 110mA is 60W dissipation, as Pentode (and maybe in UL) should give ~~24W Sine audio in 5K load.
Add 10%-20% for filtering and self-bias. (Never run a valuable bottle near 99% Pdiss in fix-bias.) 600V to 650V. (Costly cap array.)
Add 10% for G2 and driver current, 121mA. Double for stereo, 242mA.
The 120mA PT is right out.
"540V-0V-540V" gives 750VDC cap-input, 480V choke-input.
"380v-0v-380v" gives 530VDC cap-input, 340V choke-input.
None of these options are real good fits.
As a hack, 750VDC dropped to 650VDC through a 100V 242mA = 400 Ohm 24.2W (use 50W) resistor. This can give great B+ filtering with modest caps and no choke (sorry!). The added 24W of heat may not be a big deal when you have 120W of plate heat and 32W of heater-heat (total 152W) already on board. But it will need a LOT of air-space around that resistor.
This suggests hollow-state rectifier. 100V of voltage waste is not uncommon for these. Hmmmm.... 5U4 and 5AR4 are a wee bit over their ratings. PY500 is not showing any red-flag, but I never heard of it. 6AU4 also seems to work (630V out) and that one used to be super-common. Good, because it is one-diode so you need two (and a 6V 4A well-insulated winding). And I am not sure PSUD is telling me a truth. There are a billion big diodes, but if you get one which will take the surge current it may not have the drop. Of course: we are asking the rect to be a 24W resistor, and few ordinary bottles can throw 24W.
Thanks for the link Jazbo8. I had a read through. Thanks PRR for your analysis as well. I have gone through the transformers I have on PSUD2 and found, as you did, that there was no real ideal match. However, using PSUD2 I tried something that I have seen before in older equipment and schematics that I don't fully understand. That is the use of a smaller than normal capacitor and then a choke after the rectifier. This can then be followed by whatever combination of resistors and capacitors that are required to achieve the ripple reduction desired. Varying the capacitor between 2 to 6 uF varies the voltage to any level that you want. This works on PSUD2, but I have been unable to find much on using this method to adjust B+ with respect to there being any benefits or down sides. Any info that you might have on this method would be appreciated. I prefer not to use a tube rectifier because of space constraints on the chassis I have available.
I am currently playing a kt120 in SE pentode with fixed bias. My PT is about 500-0-500 and i using a dual rectifier set up into a switchable CLCLC with the first cap (5uf PIO) on a switch to make the amp LCLC. with the C1 out i have 380vdc with 2 x 5ar4. If i use c1 i have about 460vdc. I am running the kt120 at 105ma at 380vdc and around 100ma at 460v. I prefer the sound at 380v and will use that for now.
As far as using a small value cap for c1 , i have tried that and it works well and can help dial in the B+. But for the values of caps that are readily available the amount of adjustment to the voltage is limited. for example, in a different amp that i tried the small c trick, with .5uf i got just a few more volts less than 10. when i put in a 1.0uf i got over 10v. when i tried 2uf i got around 30v when i tried 5uf i got more than 50v more. the values i give are just from my memory but is more or less how the voltages increased.
As far as using a small value cap for c1 , i have tried that and it works well and can help dial in the B+. But for the values of caps that are readily available the amount of adjustment to the voltage is limited. for example, in a different amp that i tried the small c trick, with .5uf i got just a few more volts less than 10. when i put in a 1.0uf i got over 10v. when i tried 2uf i got around 30v when i tried 5uf i got more than 50v more. the values i give are just from my memory but is more or less how the voltages increased.
Thanks DAK808. I'm not sure how accurately PSUD2 will predict the use of caps of smaller sizes, and it may be a trial and error process once the amp approaches completion. I'm pretty sure it will work, I'm just not sure about any pros or cons, so I'm leery about proceeding. The great thing about doing it this way is that the only cap that needs to be a really high voltage type is the first one, and two caps of lower rated voltage can be used to achieve the higher voltage at the required uF. After that, lower voltage caps can be used. I really like the idea of making the first capacitor switchable to vary the B+, and will use it in the future. The large change in B+ that this creates will allow for different tubes to be used in one amp. If this is combined with a switchable cathode bias resistor, which I do quite often, the same amp could be used for almost every power octal. The only only thing left to deal with is varying the voltage to the preamp tubes to keep them in a good operating range. It may not be an issue, but if it is, I suppose that the cathode resistor switch on the power tubes could also be used to switch between two different cathode resistors for the preamp tubes.
About the small c1, i noticed that under 1uf the voltage is affected minimally and starts to change rapidly after that. So, by the time you reach 3uf the voltage is close to max. I recall that a value of less than 1uf for c1 helps to lower the impedance of the PS. If you can build your PS with all film caps is also an excellent procedure for a SE amp.
Have you checked out the RH Universal amp v2. It allows the front end to stay under the same operating conditions except that the voltage varies with different rectifier tubes. So going from the 5y3 or 5r4 to 5ar4 is a change of about 40 to 50 vdc. And you would apply to tubes from 6l6 to kt88 or kt90 and in my amp kt120. I also used fixed bias which allows a much broader range of cathode current and much more precise adjustments.
Have you checked out the RH Universal amp v2. It allows the front end to stay under the same operating conditions except that the voltage varies with different rectifier tubes. So going from the 5y3 or 5r4 to 5ar4 is a change of about 40 to 50 vdc. And you would apply to tubes from 6l6 to kt88 or kt90 and in my amp kt120. I also used fixed bias which allows a much broader range of cathode current and much more precise adjustments.
I usually use only film caps in my builds, except for cathode capacitors, which I normally bypass with film caps. I have built the RH84 and listen to one daily, and like it. I'll have another look at the RH Universal, but I'm likely to build my KT120 amp with triode and ultralinear connections, as my OPTs provide for both. A switch to add pentode mode with Schade style feedback might be workable too. Other than lowering impedance of the stage, are there any other impacts that you know of that result from the use of a small input capacitor?
> use of a smaller than normal capacitor and then a choke
Yes, and PSUD will show you why it works. The voltage at the first, small, cap has HUGE ripple. Say the peak hits 500V and the dip hits 300V. Now L-C filter that, the "average" is about 400V (+/- depending on exponentials, rectifier pulse-width, and choke DCR loss). With ample L and C on the end your DC may be as smooth as you wish.
The "problem" is that the first cap value is more critical. Usually the first cap "can" be -50%/+100% tolerance, and older electrolytics were certainly this sloppy, and worse as they age. (Paradoxically an e-cap's effective value can drop or *rise* toward the end of life.) Hammond did it with film caps, which can be made much better than +/-20% and don't change for many decades. But that much film cap is historically expensive. But fashionable, so maybe it is a way to go.
A hasty sim suggests the sensitivity to 1st-cap value is not large. This may be your path.
Yes, and PSUD will show you why it works. The voltage at the first, small, cap has HUGE ripple. Say the peak hits 500V and the dip hits 300V. Now L-C filter that, the "average" is about 400V (+/- depending on exponentials, rectifier pulse-width, and choke DCR loss). With ample L and C on the end your DC may be as smooth as you wish.
The "problem" is that the first cap value is more critical. Usually the first cap "can" be -50%/+100% tolerance, and older electrolytics were certainly this sloppy, and worse as they age. (Paradoxically an e-cap's effective value can drop or *rise* toward the end of life.) Hammond did it with film caps, which can be made much better than +/-20% and don't change for many decades. But that much film cap is historically expensive. But fashionable, so maybe it is a way to go.
A hasty sim suggests the sensitivity to 1st-cap value is not large. This may be your path.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.