I'm thinking about maybe building a tube amp if I'm still motivated to try another project after a speaker project. I used to know a lot of this stuff. (I taught myself enough electronics all on my own at 15yo to pass a Tech Class Amateur Radio license exam; I was a bright kid).
I read this paper to get paper to refresh my memory on theory and design.
http://www.aikenamps.com/CommonCathode.htm
And then compared the schematic's for Mikael's KT88 and the Simple SE--
http://diyaudioprojects.com/Schematics/Mikael-Abdellah-SE-KT88-Amplifier.htm http://www.tubelab.com/SimpleSE_schematic.htm
Anyone care to take a stab at the thinking-out-loud questions I have after looking over all that stuff....
Q1: SimpleSE has huge cathode bypass caps (1500uF) compared to the other. Any reason why? Once it’s big enough to effectively be an AC short, why any bigger? Maybe just at that voltage they’re cheap regardless of uF so buy what’s avaible?
Q2: Mikael’s doesn’t have an Rgrid on the driver stage. Is the volume pot functioning as the Rgrid? The 100K will pull the grid to ground no matter what the level is set.
Q3: The input pot on the SimpleSE doesn’t show a value. Will 100K work? Will the 220K change the taper? Could it be reworked to have a different Rgrid based on what the volume pot is?
Q4: The SimpleSE has 100Ohm Rs in front of both grids. The Constant Cathode Amp paper doesn’t mentions Rs in that position. What are these for? Do they affect the bias or are they just to get a little bit of isolation between the stages? I’ve heard extra passive components in the signal path can cause noise.
Q5: Mikael’s amp has a simple R on plate on the driver stage to drop the B+. The SimpleSE has drop R and then a CCS. I guess the CCS gives more linear response. I suppose it would be possible to mod Mikael’s amp to have a CCS. You have to figure out the current (easy I think by measuring V across the R drop at idle); then figure out how to set the CCS to that value (probably not hard), then probably decrease the value of the dropping resistor somewhat (not really sure how the heck to figure that out yet). Does that sound doable? I’m kind of debating which of the two I would built if I decide to build one and I like the idea of starting with the simplest design and then tweaking it. I think if I build it, my choice will depend a bit on whether I want to do P-to-P or PCB and that depends on what I can hunt down for an enclosure.
Q6: On the SimpleSE, it leaves final calculation of B+ up to the application so you have to pick a transformer which is okay. What I don’t get is the SS vs. Tube rectification and how that might mess up the calculations. I’ve played around with PSUD2 to see what transformer and rectifier, I get a much higher voltage when using SS vs. Tube rectification. I guess you have to plan so if you do put in the switch, you pick points where both modes work and if either doesn’t you just settle on one of the options before you build. Make any sense?
I read this paper to get paper to refresh my memory on theory and design.
http://www.aikenamps.com/CommonCathode.htm
And then compared the schematic's for Mikael's KT88 and the Simple SE--
http://diyaudioprojects.com/Schematics/Mikael-Abdellah-SE-KT88-Amplifier.htm http://www.tubelab.com/SimpleSE_schematic.htm
Anyone care to take a stab at the thinking-out-loud questions I have after looking over all that stuff....
Q1: SimpleSE has huge cathode bypass caps (1500uF) compared to the other. Any reason why? Once it’s big enough to effectively be an AC short, why any bigger? Maybe just at that voltage they’re cheap regardless of uF so buy what’s avaible?
Q2: Mikael’s doesn’t have an Rgrid on the driver stage. Is the volume pot functioning as the Rgrid? The 100K will pull the grid to ground no matter what the level is set.
Q3: The input pot on the SimpleSE doesn’t show a value. Will 100K work? Will the 220K change the taper? Could it be reworked to have a different Rgrid based on what the volume pot is?
Q4: The SimpleSE has 100Ohm Rs in front of both grids. The Constant Cathode Amp paper doesn’t mentions Rs in that position. What are these for? Do they affect the bias or are they just to get a little bit of isolation between the stages? I’ve heard extra passive components in the signal path can cause noise.
Q5: Mikael’s amp has a simple R on plate on the driver stage to drop the B+. The SimpleSE has drop R and then a CCS. I guess the CCS gives more linear response. I suppose it would be possible to mod Mikael’s amp to have a CCS. You have to figure out the current (easy I think by measuring V across the R drop at idle); then figure out how to set the CCS to that value (probably not hard), then probably decrease the value of the dropping resistor somewhat (not really sure how the heck to figure that out yet). Does that sound doable? I’m kind of debating which of the two I would built if I decide to build one and I like the idea of starting with the simplest design and then tweaking it. I think if I build it, my choice will depend a bit on whether I want to do P-to-P or PCB and that depends on what I can hunt down for an enclosure.
Q6: On the SimpleSE, it leaves final calculation of B+ up to the application so you have to pick a transformer which is okay. What I don’t get is the SS vs. Tube rectification and how that might mess up the calculations. I’ve played around with PSUD2 to see what transformer and rectifier, I get a much higher voltage when using SS vs. Tube rectification. I guess you have to plan so if you do put in the switch, you pick points where both modes work and if either doesn’t you just settle on one of the options before you build. Make any sense?
If capacitors were perfect, you could just calculate the value needed for an AC short at the lowest operating frequency. Electrolytic capacitors are far from perfect however. Any real capacitor has some internal resistance known as the "effective series resistance" or ESR. They also have some "effective series inductance" or ESL. These two can team up to create some resonant "non shorts" in the audio range even though the capacitance value is sufficiently large. It is generally true that the ESR continues to drop as the value is increased. The ESR is often specified in the data sheets. The ESL is not as easy to predict from the value, and is not always specified in the data sheet. There are some other non ideal properties that can affect the use of capacitors in audio. The "dissipation factor" or DF is what causes the "smearing" and loss of resolution in coupling capacitors. These properties can be measured. In this case I got some time with a Hewlet Packard component analyzer and tested a big bunch of electrolytic caps. I found that the 1500 uF Panasonic capacitors specified in the parts list had the lowest ESR over the audio range, and the ESL did not cause any resonances in the audio region. Some 150 and 220 uF capacitors exhibited resonant peaks in the audio range where the total impedance rises often to a few ohms. This may not make much difference for the bypass on the 12AT7, but it will be audible in the output stage especially if CFB is used. If you are ordering new parts, I recommend the parts in the parts list, if you have some caps already, use them.
Again if the volume pot was perfect a resistor to ground would not be needed. As the pot ages, it can develop dead spots where the wiper is intermittently open. Contrary to popular opinion allowing the grid to float momentarilly will not cause the tube to self destruct, but it will cause a nasty pop in the speakers that may not be good for them.
I would use the lowest value that your source is comfortable with. Many CD players and SS preamps are happy with 10K while some tube preamps require a 100K load. If the pot is too high in value, it can form a low pass filter with the Miller capacitance of the input tube. The NOS 12AT7 has a low capacitance, but some new production 12AT7's can be higher. I would use a 100K pot or lower. Rgrid can be any value between 100K and 1 meg. Low values can affect the "taper" of the pot.
These are "grid stopper" resistors which are standard in most tube amplifiers to reduce the possibility of oscillation. The value required is dependent on the layout and construction of the amplifier. If the value is too low the amp may oscillate on signal peaks leading to an annoying "fatiguing" type of distortion. If the value is too high the high frequencies will be rolled off. The PC board layout of the SimpleSE has been optimized to allow the use of 100 ohm resistors.
The CCS allows the 12AT7 to operate with a very high effective load impedance giving a much lower distortion and higher gain that a resistor alone can provide. There have been a few builders who have expressed the desire for a "sand free" SimpleSE, so the CCS is simply bypassed and the resistor is increased to 33K ohms. The 10K resistor is there to absorb some of the heat dissipated in the CCS and provide some isolation.
For 95% of the amplifiers being built it comes down to two choices. Use a 720 or a 750 VCT transformer like the Allied 6K7VG for $44, if you are using the usuall 6L6GC, EL34, or KT88 tubes. Use a 550 VCT transformer (Allied 6K56VG) for the 6V6 amp. The Allied transformers are made by Hammond, but cost less than the equivalent Hammond. The charts were intended for the builder who wanted something different, or wanted to use some existing components, and needed to choose the rest.
In the real amplifier I get about 430 volts with a 5AR4 and 450 volts when the SS diodes are used. The amplifier makes a little more power with the SS diodes, and has a harder sound, often with more bass. I added this option for those who want to use it. I use the SS position when cranking some Pink Floyd loud. Many user simply leave the diodes out and do not have the SS option. They are not needed for normal operation with a tube rectifier.
That is often the deciding factor since the two designs are similar. The tricks outlined above could be applied to either design. If you are comfortable with point to point construction you could build a SimpleSE from the schematics and parts lists. This has been done by some builders. The PCB makes construction of the amplifier easier, and more repeatable. All of the usuall hum, noise, oscillation and mis wiring issues have already been solved . I have received a few emails asking for help fixing hum in a hand wired SimpleSE. I can not possibly fix hum in an amplifier that I have never seen.
Thanx for the info tubelab. You've been so much help with my learning this stuff I think I'm going to have to buy something from you some time.
Speakers first but the cheaper transformer and edcor OPTs could put the SimpleSE in my price range of a few hundred $ as my next project.
So, two questions...
When are the new SimpleSE boards in stock?
What are the board dims so I can figure out an enclosure?
Speakers first but the cheaper transformer and edcor OPTs could put the SimpleSE in my price range of a few hundred $ as my next project.
So, two questions...
When are the new SimpleSE boards in stock?
What are the board dims so I can figure out an enclosure?
When I build Mikael's schematic I used a 100K grid stopper and a 4.7K grid resistor on the input stage. I did this primarily so I could build the amp as a monoblock power amp with a separate preamp. However I have found that using a volume pot of either 50K or 100K in front (as in a "passive preamp") works just fine with that layout.
The design sounds very good. That said, Tubelabs Simple SE has incredible versatility and according to all accounts I've read it also sounds great. I don't think you'll go wrong with either design.
The design sounds very good. That said, Tubelabs Simple SE has incredible versatility and according to all accounts I've read it also sounds great. I don't think you'll go wrong with either design.
tubelab.com said:
Sigh, I am definately going to have to get one and start a project on one of these little puppies. I have spent many an hour on your site reading and rereading what knowledge you have to offer. I'm overwhelmed, to say the least.
You may not be the prettiest thing out there, but your site is just beautiful, and a joy to cruise through.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.