Playing with the layout a bit, here is a view of potential layout of the amplifier board. The PCBs for mounting sockets, turrets and switches are very handy. It accelerates the build process and provides full flexibility.
You may not like the open look and feel, but who cares! I love the aesthetics and those OPTs are enormous!
Those OPTs are enormous, who makes them?
I like the ground plane (presuming that’s what it is) PCB.
A friend from Bulgaria. Will share his details after reviewing this OPT. Not sure how much bandwidth he has to wind OPTs at the moment. I will get some for an 845 in the near future (I hope)
Yes, it’s a combination of ground plane and a mounting board. I’m planning to stack 2 of them. Lower one will have regulators, top ones will have hybrid mu-follower boards, D3a (noval sockets), RCA connectors etc. The top board can be replaced for another one with UX-4 sockets or UX-5 sockets. So I can play around with different drivers. It’s all fun!
Excellent idea. I’ve been designing some similar modular boards, good excuse to learn Kicad, I really like the idea of a ground plane and mounting board in one. Very convenient!
I’m interested to read your thoughts on the OPTs. They are substantial to say the least!
PCB design is more difficult than it looks for audio work in which most of the parts are either NOS or rare. You have to design your parts in the CAD tool. That takes time. A lot of trial and error with the PCB prototypes which unfortunately leads to wasted money and time spent. Once you start getting the handle of it, then it becomes an easier job. Just like everything else.
I invested a lot of time designing this flexible small boards for this purpose. I wanted to build quick and have a flexible build to play around. I have limited time and need to make the most out of it!
Let’s see how this one turns out. I may struggle to fit everything in one chopping board, in which case will push me to a mono block approach. We’ll see.
I made the latest Mule preamp (RE-084) using these stacked PCBs and tinny boards. Amazing result, will post more pictures when I get the time so you can see how I’ve done it
Ale
Hi Ale,
Alexander makes interesting designs. 🙂
I’ll be curious about your experience.
Regards
euro21
Hi Bela
Yes, he’s taken quite a lot of time to design this pair of transformers for me based on my specifications. The measurements are outstanding.
Here are the specs:
Impedance ratio: 3.3k/8R and 3.7k/4R
Primary inductance: 24.5H
Nominal primary Idc: 80mA
Max power output at 25Hz*: 12W
Primary coil Rdc: 110R
Secondary coil Rdc 8R: 0.42R
Secondary coil Rdc 4R: 0.21R
DC flux density at 80mA Idc: 0.54T
AC flux density at 200Vrms**: 0.49T
*Theoretical power output without the losses included
**This primary AC voltage is equal to the maximum theoretical power output
Shunt capacitance: 2.3nF
Leakage inductance: 1.28mH
Cp/Ls resonant frequency: 93kHz
Cs/Ls resonant frequency: 200kHz
Frequency response (800R Zout): 5Hz-80kHz (-3dB)
And by the way Bela, as you are the “300B man” here, how do you see running the 300B at 370V/80mA? I used it before around that or slightly higher current. I Think it should give less than 8W and bias around -73 to -75V. I need to do the load lines again.
Thanks
Ale
I used my 300B tubes at -about- 380..385V, 70mA … but with 5k OPT.
370V/80mA is below the 80% safety limit, if you use metal anode tubes.
My amp produce easily 10W (the anode voltage “run out” slightly over 11W), but I use its even A2 region (about up to +20V peek), with even over 200Vpp grid swing.
If you use 300B in A1 region, the expected power is about 6-7W (depending of operating point and load).
B+: 398V
EH300B gg (-78.2V, 70.0mA, 385V)
PWR L_THD R_THD L_Ug THD R_Ug THD
1W 0.35% 0.31% 22V 0.11% 21V 0.12%
4W 0.86% 0.80% 47V 0.31% 45V 0.31%
6W 1.23% 1.08% 57V 0.39% 54V 0.38%
8W 1.52% 1.37% 65V 0.43% 63V 0.45%
9W 1.56% 1.43% 69V 0.43% 66V 0.45%
10W 1.57% 1.47% 73V 0.40% 70V 0.45%
12W 2.60% 2.14% 83V 0.25% 79V 0.40%
Great feedback Béla. Thanks. I will allow this to operate in A2 as have the SF with +50V and -200V. I don’t think I’d ever need 10W though. With my current 4P1L PSE I’m more than ok with volume. Sometimes could use with a tad more when I’m enjoying music with the kids around. So I think I could do with 5 or 6W maximum. Nevertheless, the circuit will provide enough swing in the SF stage.
I can easily drive down to 385/70mA. I will get on the load lines.
I have several pairs of Chinese , EHs and a pair of beloved EmissionLabs.
Nice to see you’re making progress, Ale. Keep up the good work
I’m working on the 845 OPTs though. The first pair for our friend Rob, of course. Little overdone on the LF though, 3Hz to 50kHz, 35W output.
Best regards!
Hi Alexander,
Thank you, hope to finish this amp soon. I’m keen to test the OPTs and provide feedback. I have quite a few around for the 300B
Rob has been keeping me up to date with the progress on the 845 OPTs. I’m keen to get mine as well, as the 845 will be my next project after this one!
Keep me posted with your progress as well
Cheers
Ale
Hi Ale,
Let me know if you’re in a hurry about the 845 OPTs, so that I systematize my other projects as well.
I guess you wish to finish this one first. 🙂
Good luck on the build! I like that prototype module freedom by the way, I started making my tube sockets with output pins.
Best regards,
Alexander.
What are the dimensions of the output transformer?
Hello Martin,
If you’re asking about Ale’s 300B nano OPTs, they’re – 130*125*96mm
Best regards!
Thanks Alexander.
Did you use the amnc-120 core or even nicer, a costum core?
Specs look very good btw, just the copper losses a bit high @ o,36dB
Hi Martin,
The core is a regular standard AMCC320. The specs can become even better, if the core gets bigger. Nanocrystalline and amorphous are bothersome materials to work with, because of the maximum flux density of 1T* and the stacking factor of 0.78.
This is why you have to get much bigger than HiB to compensate.
P.S. Althoug amorphous is rated at 1.56T and nano at 1.27T max flux density, both start the severe saturation knee at 1.0T
Oh, nice core. That would limit the turns a lot to about 2000-2100.
I also noticed that the 1,56T of the Amorphous iron material is a bit exaggerated. Maybe that’s why many amorphous transformers on the market have a bit weak bass.
The primary number turns of Ale’s transformer is on the 4000 side. This is a hint how hard is to squeeze basic parameters of nanocrystalline cores.
From experience, weak bass can be due to the wrong combination of materials. This is probably the most audible, but not measurable parameter of an audio transformer. Many today’s manufacturers seem to neglect this, but I suspect the legendary brands such as Tamura were aware.
With 4000 turns, B=0,49T /200V i calculated 12,7Hz.
Maybe i calculated wrong because L=24,5 H . What am i missing?
Here you go:
Bdc = Idc * L / Afe * N
Bdc = 0.08 * 24.5 / 0.0009 * 4080
Bdc = 0.53T
At 200V, 25Hz
Bac = 200 / Afe * 4.44 * F * N
Bac = 200 / 0.009 * 4.44 * 25 * 4080
Bac = 0.49T
Btot = Bdc + Bdc = 1.03T
“No” Alexander.
If you say that you use a amcc320 size core and the photo on Ale’s website shows clearly that there are 2 cores inside then Afe should be 0,0018m^2 and not 0,0009m^2.
The amc320 has a “stack” of 50mm and that won’t match with the outer dimension of the transformer i think. With 2 amcc320 i think it will match.
My calculation was with Afe= 18cm^2 so probably we are still on the same road but we disagree on the core dimensions.
In case it is realy a core of just 9cm^2 you are right.
But, that means its it’s just a small transformer for the job……
You are mistaken by the core’s specific economical construction.
It is an equivalent AMCC320 size fused by two pieces. This is how the manufacturer makes them. It seems to be a recent, economical way of manufacture in order to avoid scrap losses.
The core’s Afe is truly 9sq.cm and yes, it is small and building an output transformer with is a harder task.
The primary idea was to manufacture a more budgetary kind of nanocrystalline transformer for Ale.
Wow, that’s new to me. Then you did a hell of a job because it ain’t easy to wind 4000 turn and still have 80kHz and 5Hz (Ri=800 Ohm) low end.
Is the fase response also nice? As Tango dos?
There are tricks to get away with more turns, but let’s consider this project to a squeeze to the maximum. It will be really hard to do better.
High frequency phase response is dependant of the resultant RLC filter type you achieve with your parasitic components. Usually I get smooth phase roll-off, between Bessel and Butterworth type. But I can get any phase response I want. Again, I believe the sound of the transformer is more hidden into its materials and combination of materials than electrical parameters.
For me it’s simple, you need both. Good materials and electric parameters.
Ale,
I’ve been trying various drivers for my 300b amps. Which are set up in auto bias. I am not an expert at this, but enjoy trying things and making my system better. I’ve been at it now 30 years. I’m gonna try the d3a next. What would be the easiest way to drop it in? Use you follower pcb’s?
I almost forgot, I wanted to ask/encourage you to to make the various pcb’s you used available for sale. They may seem a trivial matter, but how many times am I gonna cludge together some of the basiccvbuilding blocks, the driver heaters to name one. I’d buy most of the ones you’re using in this build.
Hi David
Not sure if you’ve read my latest blog post. If not, let me repeat that implementing a DC coupled amp is a complex and challenging build. I don’t want to put anyone off, however I want to give my honest opinion after implementing several of them.
D3a as a driver is superb. In my design, I get about 3Mhz of flat response at 37-38dB which is impressive for this driver. Also it has the lowest distortion out there when swinging +150Vpp. I measured all sort of different drivers, and in my experience the D3a wins. It sounds great. If you are not getting into the complexity of DC coupling, you can then try cap coupling the D3a into the 300B grid. The D3a can drive the grid without a problem.
With regards to the PCB request, I need to find some time. I made a few of them in small batches, so the cost isn’t that low when you factor in the development time, prototyping costs, shipping and taxes.
I will make them available when I can as I know they can simplify life like hell when you’re building audio gear. In particular, I have various turret boards made from 1 to 4 and single/double lines. Great to locate and place components around for point to point wiring.
Hope this helps, not sure how much interest will be on these boards but will let go some of my stock to whom is interested in them.
Cheers
Ale
Hi David
I just added a page with some building block PCBs for sale under the For Sale section.
Cheers
Ale