If you wonder what the hoo-haa about the 1545 Non-OS DAC is about, please refresh yourself by reading my impression here, the "Monica" DAC project.
Perhaps you want to read some customers' reviews?
Shown above is the blank PCB. At this point, I should be showing you the schematic, but I'm too lazy to draw it up. Anyway, it's not much different from the "Monica" page. Besides, from a layman's perspective, what the heck about schematics? Looks like spaghetti to me! So here's a much easier assembly guide! (pardon the long download time). Click on image for higher resolution (and even longer download time). If using Internet Explorer, press "F11" for full screen view.
Is the above assembly guide self-explanatory? I certainly hope so. I'll have a FAQ section later. Note that C3, C4 and C10 are optional. Solder 0.1uF to bypass the big electrolytics.
There are 2 power supply connections, J4 and J5, for those who might want to experiment with dual supplies. If using just one supply, wire round via of J4 to round via of J5. Solder incoming supply to J5.
| 1545 Non-OS DAC Parts List | ||
| U1 | CS8412CP | |
| U2 | TDA1545 or TDA1545A | |
| U3, U5 | 7805 | |
| U4 | LM317 | |
| C2, C5 | 47uF or higher, >6.3V | |
| C3, C4, C10 | 0.1uF. >10V, Optional | |
| C6 | >10uF, >10V | |
| C9, C16 | >2uF, >10V | |
| C11 | >10uF, >10V | |
| C14, C22 | 0.1uF, >10V | |
| C20 | 3300pF, >6.3V | |
| C21 | 0.22uF, >6.3V | |
| R16 | 75 ohm, 1/8W | |
| R12 | 1kohm, 1/8W. Optional. Install with LED D4 or omit both. | |
| R17 | 470 ohm, 1/8W | |
| R18 | 33kohm, 1/8W | |
| R19, R20 | 2-2.2kohm, 1/8W | |
| L1, L2 | >1mH or just replace with wire | |
| D4 | LED. Anode (longer lead) goes to square via. Optional | |
| D10-D18 | Any normal diode will do, like 4001, 4007 etc. | |
About choice of diode. Man, I had a lot of fun with this. By right, any diode will do but since we are relying on the forward voltage drop of a diode, which varies from diode to diode, this means we'll have different VDDs with different diodes!
I built a prototype using some Schottkies (it was a Sunday and the only 9 diodes I have were Schottkies) and the string of 9 diodes came up to 5.6V, just a little lower than the Absolute Maximum rating of 6V for TDA1545/A. This PCB is designed for 1N4148 so I wired in 9x 1N4148 but the VDD came up to 7.6V! This is TOO HIGH!
Strangely though, my TDA1545 is still singing. The chip doesn't even get warm and I can't hear/detect any distortion either. So what's the problem?
Problem is I can't guarantee every TDA1545/A will behave in the same manner and (I can't guarantee my behaviour as well). Philips explicitly said that the Abs Max is 6V so going beyond that is at your own risk! So to be on the safe side (it's less fun but we can't live dangerously all the time), please build it with a string of 6 diodes. Replace the 3 with wires.
So you law abiding Diyers, substitute D12, D15 and D18 with pieces of wire. Just short them. Use ferrite beads if you are macho.
You outlaw Diyers, go the full 9 diodes and pray!
I built the DAC as above with just 6 1N4148 diodes and shorting D12, D15 and D18. VDD came up to a nice 5V.
Building the DAC TDA1545/TDA1545A Non-OS DAC PCB. Double sided, FR4, with solder mask and silkscreen. USD25 or RM90 each. I accept PayPal!
TDA1545A DAC chips. USD3.50 or RM12 each.
You provide battery, charger, solder, wire, RCA sockets, casing and plenty of attitude!
Update on Aug 17th
This should be straight forward. Follow assembly guide above.
Solder everything in except CS8412 and TDA chip.
Wire or hook up power supply. Measure across power supply caps - C2-C5 (5V), C10-C11(~4.2-6V, this is your VDD).
For TDA1545/A, measure also voltage at +ve pin of C6, or the C6 end of R18. This should read 2/3 of your VDD.
Once satisfied voltage above is correct, solder in the chips.
Hook up to CD/DVD player. "Play". If LED lights up, this means that CS8412 managed to "lock" onto the signal. You are ready to play!
If you didn't wire in LED, measure voltage on pin 28 of CS8412. It should read ~0V when "locked", otherwise ~5V.
Ha! This is where you give me money!
"Enhanced" TDA1545 Non-OS DAC PCB. Above PCB but with 1.2mH inductors, ferrite bead and 47uF Oscons. USD30 or RM110 each. (picture shown is of 1543 pcb)
Full kit. USD85 or RM320 each. Assembled and tested for you, USD115 or RM420.
Here's what you'll get.
CS8412CP digital receiver IC
TDA1545A
Motorola 7805
On Semi 317
Low-profile 28 pin socket
Elna Cerafine 10uF output coupling capacitor. Add RM10/USD4 for Black Gates 10uF ala above.
Elna Cerafine 10uF power supply decoupling capacitor
Sanyo Oscon SP 47uF
Evox Rifa 0.1uF
Evox Rifa 0.22uF
Wima 3300pF
1.2mH inductor
Ferrite bead
Blue LED
1% Metal film resistors, 75ohm, 470ohm, 1kohm, 2.2kohm, 33kohm.
Oh boy! If I have my way, I would like to have a word with Philips product engineers! I was going through TDA1545 datasheet this morning when I noticed the specs on page 9.
See that cryptic "R3 = R4 = 11kohm" line? THD is lower by 0.4% while SNR is higher by 3dB. Huh?
R3 is internal Rreference in the DAC and specified as typically 11kohm. So setting R4 to 11kohm as well means that we are setting Vref at 1/2 of VDD instead of datasheet's recommendation of 2/3VDD.
So I had my DAC done up as above. Replaced R18 with 11kohm, wired 1N4148 diodes for D10-D15, shorting D16-D18.
Played music and... Guess what? There is less warmth in the presentation. Seems cleaner as well. Nothing else changed but the signal output is way higher now. My volume control does not even need to travel much. Of course, if you try this mod, let me know your impressions!
So if you prefer more warmth, build for R18=33kohm and Vref=2/3VDD, though you get lower output. If you prefer a cleaner presentation, less warmth, but higher output, try this.
Wire 6 diodes but short the last 3 diodes. Or just jumper the last 3 to GND like the above. Now your Vref is at 1/2 VDD. Also, there is the issue with R18...
Philips state that Rreference is between 7.4kohm and 14.6kohm with a "typical" 11kohm. What a wide variation! Whipped out my multimeter, measured between pin 7 (IREF) and pin 4 (GND), I found some samples at 7.6kohm, some at 8kohm but so far none at 11kohm. So much for "typical"...
I guess if you are really a "sucker" for maximum dynamic range, measure your Rreference value and put in R18 with equivalent value, otherwise 11kohm will be okay as a "typical" value. This is despite Philips datasheet saying "maximum dynamic range observed with VREF=2/3VDD". Go figure.
Anyway, so far I observed that all 1545 DAC circuits around the Internet use 33kohm for Rref and set Vref to 2/3VDD. Haven't seen any other variation. (If you do, let me know!) I guess we blindly followed Philips' datasheet! But in my honest/humble/humbled opinion, 11kohm and 1/2VDD is where this DAC sings.
Another update!
Apparently some folks found 11k&1/2VDD to have some distortion in the lower bass region. I can't hear it in mine but some could. Further experimentation sees an optimum of 22kohm is just nice. No distortion at all and still sounds as nice. So all kits will be shipped with 22kohm in the future.