While searching online today for how much mW I actually need for normal listening and the maximum SPL through my Denon AH-D7200 headphones, I came across an interesting table.
Based on this, I decided that setting the bias to about 110 mA would be sufficient, so I chose R10 (10R). I have omitted the large dimensional heat sink, and instead a 10mm thick plate with dimensions 160mm x 80mm will be used as the heatsink. This plate which will serve as the main heatsink will be screwed to the aluminum bottom of the Modushop SlimLine 2U case. I drew the board in a CAD program and sent it to the factory today, it will be made by the Slovak company QuickPanel. The plate comes already pre-drilled with M3 screws in black anodized finish.
When my UltraBiB PCBs arrive, I will add the same heat dissipation mechanism to them. 😀
Based on this, I decided that setting the bias to about 110 mA would be sufficient, so I chose R10 (10R). I have omitted the large dimensional heat sink, and instead a 10mm thick plate with dimensions 160mm x 80mm will be used as the heatsink. This plate which will serve as the main heatsink will be screwed to the aluminum bottom of the Modushop SlimLine 2U case. I drew the board in a CAD program and sent it to the factory today, it will be made by the Slovak company QuickPanel. The plate comes already pre-drilled with M3 screws in black anodized finish.
When my UltraBiB PCBs arrive, I will add the same heat dissipation mechanism to them. 😀
For the DCG3 mosfets alone I think it will work and it has been done by others too but.... are you sure that this will suffice for UBIBs for normal weather conditions ?
I'm always skeptical about mounting big heat transmitting devices to the bottom of the chassis. I have a similar arrangement for my UBIBs and at the summertime is not optimal and I'm in the process of changing that, even relocating to a new custom chassis
I'm always skeptical about mounting big heat transmitting devices to the bottom of the chassis. I have a similar arrangement for my UBIBs and at the summertime is not optimal and I'm in the process of changing that, even relocating to a new custom chassis
Can you post a detail photo of how you are currently cooling the UBiB, since it's so hot? For the UBiB, I plan to use a single 10 mm x 128 mm x 185 mm plate for both dual-mono PSU.
I have dismantled DCG3 but all my four UBIBs were standing (mosfets in horizontal position) on top of 160mm x 300mm plate/heatsink 15mm thick
I have two 19VAC EI core transformers which provide 23VDC so no extra energy needed to be dissipated via heatsink.
Yet the warm months it gets too hot.
In your application the bottom plate (3mm thick) is gonna act as additional heatsink for the preamp but also for the UBIBs. Remember it's a class A design and UBIBs will have to provide extra current.
I have two 19VAC EI core transformers which provide 23VDC so no extra energy needed to be dissipated via heatsink.
Yet the warm months it gets too hot.
In your application the bottom plate (3mm thick) is gonna act as additional heatsink for the preamp but also for the UBIBs. Remember it's a class A design and UBIBs will have to provide extra current.
I've used UBIBs in my DAC. In a particular position, there was no room for big caps and heatsinks. You can see them with the mosfets mounted on small green PCB leftovers mostly for support. R1=3,9 ohm so CCA=150mA and Vout=8V. The chips roughly consume 50mA. I got it running cool using preregulators -linear LM7812. According to UBIB manual, even smaller Vin-Vout margin could work.
Attachments
That reminds me of a DAC I've build for a friend of mine, 5 BIBs and 2 or 3 Reflektors. In the end my friend sold it and his biggest concern was the heat.
@MagicBus, I thought I was the only crazy one.
Some facts:
UBIBs are more efficient: less current wasted and less voltage headroom needed
DCG3 is class A and needs extra current so for 100mA consumption, I set the CCS to 250mA. Take note that negative side consumption is much less.
DACs usually need 5VDC or so while DCG3 needs 17VDC
Total consumption for mine is 39 watts including of ~4W for muse attenuator and I-Select
@MagicBus, I thought I was the only crazy one.
Some facts:
UBIBs are more efficient: less current wasted and less voltage headroom needed
DCG3 is class A and needs extra current so for 100mA consumption, I set the CCS to 250mA. Take note that negative side consumption is much less.
DACs usually need 5VDC or so while DCG3 needs 17VDC
Total consumption for mine is 39 watts including of ~4W for muse attenuator and I-Select
Diy fever is pandemic... 🙂
My DAC gets moderately hot. No problem even at summertime. I tried to optimize supply/consumption ratio. And for a reflector-D that provides 5V/450mA out of a 9V transformer, I put bigger heatsinks.
I'm not sure I get it right, I think that it's not the final output voltage that loads the mosfets in UBIB but the input/output difference. That's how I size everything. So, for a 18V AC transformer to 17V/200mA DC it should be more than 1,6W given the unsmoothed input. But for say 20V regulated DC to 17V/200mA it's only 0,6W.
Unfortunately, even this cannot help with my doubled DCG3, that's why I stay with DCSTB.
My DAC gets moderately hot. No problem even at summertime. I tried to optimize supply/consumption ratio. And for a reflector-D that provides 5V/450mA out of a 9V transformer, I put bigger heatsinks.
I'm not sure I get it right, I think that it's not the final output voltage that loads the mosfets in UBIB but the input/output difference. That's how I size everything. So, for a 18V AC transformer to 17V/200mA DC it should be more than 1,6W given the unsmoothed input. But for say 20V regulated DC to 17V/200mA it's only 0,6W.
Unfortunately, even this cannot help with my doubled DCG3, that's why I stay with DCSTB.
In-Out Vdif and CCS current level are central to heat management because M1 runs at full constant current while M2 shares some with the load.
Salas, in my case on M1 0,3A/ 25V DC, the transformer is 18V AC (7,5VA loss) and on M2 0,12A / 17V DC (2VA loss). I ordered a quotation for the ZH 7505 heatsink in black elox from a Slovak company. The heatsink will be 128 mm long. The heasink has Rth at 100mm 3,21°C/W, will it be enough?
4x ZH 7505 for dual mono UBib
4x ZH 7505 for dual mono UBib
Well I have been waiting to start making posts on diyaudio until I was ready to start my project that I would actually take criticism of my work in kind. So I have just confirmed with one of the legends of diyaudio Tea-bag that my (1 of 2 separate ) DCG-3’s with Ubib power supply has shipped! So going forward and to make clear my plans are to configure and build 2 separate balanced DCG-3 pre amplifiers. One for the Left channel and one for the Right! 4 DCG-3 boards in total! Any advice and ideas are welcome!
I am also tossing around the idea of integration of a PLC to run the complete unit (just as design ideas) removing more from the signal path in the end. For the purposes of start up control volume attenuation and R1 switching to for variable input and output impedance to accommodate different input scenarios and multiple parallel power amplification and an active crossover interface. But I am going to start the process of basic assembly of PSU’s and original circuit designers (Salas,Tea-bag, and other diyaudio peers) implementations! So as my mind keeps flying I have to get rid of my last project which is sadly not audio.
The DCG-3 has been the best pre-amp/head-amp I have had in my system. I was running an Academy Audio Muses volume control, as the attenuator, but static electricity, a real problem in our cold, dry Canadian winters, messed up the controller. I replaced it with a made in china stepped attenuator but it felt like the magic was gone. Grainy and colored. I had a pair of the old $200 Intact Audio autoformers tossed aside that didn't work out in another project so I connected these up to the DCG-3. The MAGIC IS BACK. The muses volume control was last used a few months ago, so my audio memory may not be that great, but the autoformers seem even more clear, like looking through clearer water whereas I didn't realize the water was slightly murky before. I do have a question however, should I be loading the input of the DCG-3 with a resistor, say 10k across the autoformer output?
R1-->C1//R2 do present an input load already, although a light one but almost constant until very high frequencies above the audioband