I’m going to start using the 35V rail with the same 159z chokes, just experimenting with operating points. The next phase will probably be trying larger chokes (like 193v) and thf51’s. As Ben mentioned, my first step is actually replacing PSU caps to accommodate more voltage. Same for output cap.
One nice thing about my specific Antek 28v toroid (AS-4428) is that it also has two 15v secondaries, which will make it easy to whip up a couple of small circuits for manual biasing.
One nice thing about my specific Antek 28v toroid (AS-4428) is that it also has two 15v secondaries, which will make it easy to whip up a couple of small circuits for manual biasing.
Raising the supply will inch you up the Vgs curve and give more bias, but it also result in more dissipation. So watch for that. As an example, 25V and 2A gives about 50W on the device, but 50V and 2A gives 100W, probably too much for the 2SK2087c, but maybe on the edge of crazy for THF-51 and the like.
A little higher voltage around 30-40V should be fun to try.
A little higher voltage around 30-40V should be fun to try.
Yup, that sound wise.
Another thing to mention is that you will clip either with the bias current or with the supply rail plus whatever the choke can do. So, while more voltage can give you nicer performance, usually, you only want enough voltage to not clip. I am finding that those limits are not as easy to calculate. For example, at 2A bias, I get mild clipping at 13W out, which is about 30V peak to peak, higher than the supply voltage, helped along by the choke.
Another thing to mention is that you will clip either with the bias current or with the supply rail plus whatever the choke can do. So, while more voltage can give you nicer performance, usually, you only want enough voltage to not clip. I am finding that those limits are not as easy to calculate. For example, at 2A bias, I get mild clipping at 13W out, which is about 30V peak to peak, higher than the supply voltage, helped along by the choke.
Thanks. This is the simplest solution for me, parts-on-hand wise. My 10mF output coupling caps were spec'd at 50V with this in mind.
Here's a way to get higher bias from the 193V if you cannot get to your bias current target. Using Cody's PCB and schematic, the change to make is to connect R1 to the bottom of R3 and R4. The parallel resistance of R3 and R4 should be calculated solely to provide the needed bias.
For example, let's assume that no resistance is used in R3 and R4 and you are using the big 193V choke with a DCR of 1.0R. The bias current you measure for your device comes out to 2.0A. You want to get to 2.5A, which means you'd want a smaller resistance than the 1.0R provided by the choke to create a smaller negative Vgs. The best option is to get a choke that can do higher current and has a low enough DCR. Absent that and assuming you are using the 193V, here's what to do: connect, say, 0.75R in the R3/R4 position, and instead of connecting the bottom of R1 to ground, connect it to the bottom of R3/R4. Then the Vgs is solely the voltage across R3/R4, which, given that the DCR is < 1.0, would then give a bias current >2.0A. Some experimentation may be needed. Watch the power dissipated by R3/R4 and use double that for the resistor, at a minimum.
Inserting R3/R4 results in a slight reduction in the voltage swing available, but given that we are talking about only a volt or two, it is not a bad trade-off, especially if you want higher power and less distortion. You guys have talked about operating it with higher voltages, which should also help.
So, in the end, we are not adding new part, just finding a new connection for R1 and that does the trick! May the force be with you and may you be blessed with SITs that hit your desired bias targets right from the off
I am listening to the circuit below with R1 to ground, nothing in R3 and R4, and the 193V. My two channels are at 2.2A and 2.6A and it sounds fabulous.
For example, let's assume that no resistance is used in R3 and R4 and you are using the big 193V choke with a DCR of 1.0R. The bias current you measure for your device comes out to 2.0A. You want to get to 2.5A, which means you'd want a smaller resistance than the 1.0R provided by the choke to create a smaller negative Vgs. The best option is to get a choke that can do higher current and has a low enough DCR. Absent that and assuming you are using the 193V, here's what to do: connect, say, 0.75R in the R3/R4 position, and instead of connecting the bottom of R1 to ground, connect it to the bottom of R3/R4. Then the Vgs is solely the voltage across R3/R4, which, given that the DCR is < 1.0, would then give a bias current >2.0A. Some experimentation may be needed. Watch the power dissipated by R3/R4 and use double that for the resistor, at a minimum.
Inserting R3/R4 results in a slight reduction in the voltage swing available, but given that we are talking about only a volt or two, it is not a bad trade-off, especially if you want higher power and less distortion. You guys have talked about operating it with higher voltages, which should also help.
So, in the end, we are not adding new part, just finding a new connection for R1 and that does the trick! May the force be with you and may you be blessed with SITs that hit your desired bias targets right from the off
I am listening to the circuit below with R1 to ground, nothing in R3 and R4, and the 193V. My two channels are at 2.2A and 2.6A and it sounds fabulous.
Thank you, @ra7. I love the low part count and the feedback you have received.
I'm guessing the key Tokin part could be hard to find, but the rest seems obtainable.
There are lots of First Watt supply board out there too.
I am more concerned about the proper steps to bring it up and setting the bias in a manner that will stick.
TDV Journey, round two. I popped in a new transformer with 28V secondaries, bringing my final, loaded rail voltage to 37V5 (also replaced PSU caps to accommodate voltage increase). I kept my R3 resistor at 1R, which is what I'd been using for the previous 27V5 rails. Iq went from 1A4 to 1A6. Tmrw, I plan to start lowering resistance to see how it sounds/measures around 1A8 Iq. Total THD fell from .065% (27V5) to .018% (37V5). Serious cancellation of 2nd happening at this operating point – amp sounds clean and powerful, with loads of great bass. I'll report in the next day or two when I play around with Iq more.
Attachments
Yeah, if you have any of the Tokin SITs, this topology should work because they need less than -3V of Vgs to get it in the right ballpark. The Sony parts need more Vgs at which point you begin wasting too much supply voltage. But the THF-51, 2SK180, and of course the 2SK2087 are all fine.
The bias is the simplest in this amp. If you have built amps before, this should be easy. If you haven't, some amount of caution is required. But it is still relatively easy because there are no pots to turn. If you tell us your choke, we can tell you approximately the resistance needed for a given bias current. Lots of folks here who have built this one and others that can help.
Haven’t tried anything higher than the 1A6 yet - hope to take it close to 2A today. I’m still working with the 159zc, so limited to 2A at the moment. Still need to buy a couple of the bigger Hammond, but that’ll also mean moving into a bigger chassis (or seriously rearranging my current chassis). I’ll probably try THF51 at that point. That’ll be round three
I played around with different Iq values a bit this afternoon. First I tried .75R at R3 (1.5||1.5), which gave me right at 2A Iq. After measuring and listening for a bit, I tossed in a 2R for one of the 1.5R, giving me a value of .86 at R3. This brought me to 1A8 Iq, which I think is the best of all the operating points I've tried. H2 is dominant by 14db or so, and distortion is still reasonably low. I plan to try THF51 and a bigger choke at some point in the future, but for now I'm calling my TDV 'done.' That said, my transformer has two unused 15V secondaries, so may try external bias at some point. LOVE THIS AMP!
Final values: 1A8 Iq; 37V5 rails; .86R at R3
Choke: Hammond 159zc (2A/60mH/.7DCR)
Transformer: Antek AS-4428
Final values: 1A8 Iq; 37V5 rails; .86R at R3
Choke: Hammond 159zc (2A/60mH/.7DCR)
Transformer: Antek AS-4428
