After few tests and revisions (Rev 04) I updated the capacitor multiplier board I developed time ago
Despite you can place WIMA DC-Link film caps across the entire board, the use of high-quality electrolytic capacitors for the input and multiplier places is more efficient and convenient. I leave the expensive WIMA for the last cap after the SiC Mosfet:
The board has the flexibility to feed C2 with either the CCS or a simple resistor divider from the input HT. With the CCS you get above 120dB rejection of noise across the band average which is fantastic. However, there is no free lunch as you are obliged to drop 30 or 50V between the CCS for proper operation and good performance (HF in particular). The toll is carried on by the pass element (M5 or T5) which will have to dissipate the voltage drop across its shoulders.
The pass element is C3M0280090 which has a gm of 3.6S @3.5A or about 430mS @ 100mA. It also has very low Crss and output capacitance (Coss) which is ideal for good frequency response and filtering of HF coming from mains or raw supply.
I also added the feature of HV indication via an INS-1 neon bulb which is soldered straight on top of the PCB. Very handy to indicate normal operation.
With the 500k trimpot, I can regulate well a very stable voltage between 0 and 450V – won’t get close to that as it’s the limit of the electro caps and current protection ZTX458 transistor.
I’m using this board in my headphone amplifier. It’s very nice to have the slow current turn-on feature of this capacitor multiplier. With a 47uF capacitor multiplier you get above 120dB rejection. No point to increase the value of it.
I measured 300uV of noise at 250V for 50mA load. That’s -118dB which is pretty close to the Spice simulation. Extremely good!
My headphone amp is amazingly dead quiet with this HT multiplier. A great addition to it.
I made a couple of boards extra which I’m prepared to let go. Drop me an email if you’re interested.
Happy New Year Ale, I would be interested in a pair of your Cap Multiplier boards (Rev04) pcbs—with shipping to NZ. best, tim
Hi Tim.
Happy NY too!.
I have a pair of completed and tested boards I will let go. Also just a pair of PCBs but won’t provide any documentation nor BOM. I know you’re experienced so you can get on with them easily, but for other builder who asked for help in designing a cap multiplier or selection of components, I’m afraid I don’t have the time these days to assist. Hope everyone can understand.
I will ping you an email.
Cheers
Ale
Great work!
Vishay has an amazing line of capacitors 1848S/C , they are a great alternative for WIMA in this aplication, take a look!
Thanks Michal. I’ve used the 1848S in the past. I found them to be as good but they’re not cheaper (depending on capacitance and voltage of course). Some of the DC Link from Wima are very good for the price (circa £6-8).
Hi Ale
Do you still have a couple of these Cap Multiplier boards? If so, it would be great if I could order them from you (shipping to Switzerland).
Best regards
Ian
Hi Ian, drop me an email please?
Hi Ale,
This looks great!
Would there be any problem if one uses a Cap Multiplier for the B+, then puts a Constant Current Source between the Cap Multiplier and the Plate of a tube? I’m wondering if the CCS would fight against the Cap Multiplier.
Thanks,
Dan
Hi Dan
Not an issue if implemented properly. The Cap Multiplier will set a constant voltage at its output and the CCS will set a fixed current at its output. Depending on the load you have what it will vary is the output voltage of the CCS as it will push to drive a fixed current. In your example, the anode voltage will vary in line with the fixed anode current. What you need to ensure though is enough voltage drop for the CCS to operate well. At least 25V for a MOSFET-based one.
However, the CCS will present a high impedance node to the Cap Multiplier. Depending on the length of the cable and layout it could lead to oscillation. Best is to add a small film cap right before the CCS to prevent anything to happen.
Despite the Cap Multiplier has a very low output impedance and a output capacitor in the PCB, the film cap must be placed close to the CCS right at the board. Same as a source follower which can oscillate without the capacitor as it becomes a Colpitts oscillator. The SF may operate as a common gate amplifier and the leakage capacitances will provide the feedback loop needed for VHF oscillation.
A DIY-audio enthusiast (Brian F from the US) recently commissioned me a cap multiplier board for his 01a preamp based out of SiC filament bias and output transformer. It made it tricky for him to set the operating point without adjusting the HT, so the cap multiplier was very handy. Brian is delighted with his system now and sent me a few pictured which I’m attaching here.
Perfect! Thank you for the explanation.
cheers,
Dan
Feedback from Lorenz Lansky as was struggling to comment on the blog for some reason:
Some great feedback from Martjin on the Cap Multipliers. Check out the beautiful amp he built:
Hello Ale,
I’ve been using one of your Cap Multiplier boards in the power supply of a 1626 Darling amp. It makes a very quiet PS. I’m wondering, if it’s possible to connect a Cap Multiplier board to make a negative bias supply for a 6V6 grid. Do you see any issues connecting the HT Output to ground and using the earth connection as a negative output?
Thanks,
Dan
Hi Dan, I’m glad you like it. Yes, it’s a fantastic board to provide the lowest HT noise whilst minimising interaction with the amplifier hence it won’t impact as much the sound of your amplifier compared to other filtering arrangements. This board is for positive voltage filtering, not negative. It won’t work am afraid. For grid bias there are other solutions (e.g. Rod Coleman provides an excellent fixed bias regulator) which you should look into.
If you really want to use the cap multiplier as-is, you need a floating PS to feed the cap multiplier board and then ground the output whilst the GND output becomes the negative regulated voltage. Not ideal, but works.