You must be using a 32 then? I'm using the 31.
OneyedK wrote:
"RCEZ32 and RCEZ31 have different output ports, RCEZ32 is capable of more output current and has less internal resistance.
So i'd say;
RCEZ32: 392 / 93.1 Ohm
RCEZ31: 291 / 100 Ohm (rated output current is 12mA)
Bad impedance matching or overloading the output will induce jitter and heavily influence the sound."
Which upgrades do you plan on doing first?
OneyedK wrote:
"RCEZ32 and RCEZ31 have different output ports, RCEZ32 is capable of more output current and has less internal resistance.
So i'd say;
RCEZ32: 392 / 93.1 Ohm
RCEZ31: 291 / 100 Ohm (rated output current is 12mA)
Bad impedance matching or overloading the output will induce jitter and heavily influence the sound."
Which upgrades do you plan on doing first?
johnm said:
Looking at the specs for the LA9241 and LA9242 I see no difference in how the dout port is specified so I would assume the same for both. Based on accepted practice that would typically be 390 and 93 ohms or thereabout. The key is a thevenin source impedance of 75 ohms to the cable, and not to significantly exceed 12mA current from the Dout port of the DSP.
Yes I know about the( presumed) differences between the 31 and 32 output ,but yesterday I tried the 392 / 93 and I am very very iinterested with the sound .In a relaxed way . So I will listen for a week or so like that and then will try again the 291/100. And I am quite amazed hearing such differences .
Ok my dac uses CS8416 at 3,3 supply and from the datasheet the inputs accepts very little currents . I mean at this little levels I have the suspect that it can be done even better on power supply and caps . Now I am using 1% metal film resistors for the divider , and I am planning to find some carbon .
I have some specific idea for supporting frame and when realized will let you know ( and see maybe ehehe )
Ok my dac uses CS8416 at 3,3 supply and from the datasheet the inputs accepts very little currents . I mean at this little levels I have the suspect that it can be done even better on power supply and caps . Now I am using 1% metal film resistors for the divider , and I am planning to find some carbon .
I have some specific idea for supporting frame and when realized will let you know ( and see maybe ehehe )
sorry for the delay. Yes, the visions.ca units are legit. Also, Best Buy (Canada) has these in stock from time-to-time. I was able to get one from both sources. Looks like the stock from both vendors is well into the 100s of units.
As for a group buy... IMHO, the only way to do this would be someone in Canada to organize the whole shebang, or ship the disassembled units in a single shipment to someone who would then organize the group buy.
If we did do a group buy, I'd recommend having a parts option as well, as I ended up ordering all the parts from 3 separate vendors. This option would be for those who don't really have a large stock of parts to tweak with, or just want to go with Peter's recommendations.
On another note, I recently ordered a custom R-Core 2x10 VAC 50VA transformer that I'll be able to compare to the Hammond EI that Peter recommends. It won't be here for a month or so.
As for a group buy... IMHO, the only way to do this would be someone in Canada to organize the whole shebang, or ship the disassembled units in a single shipment to someone who would then organize the group buy.
If we did do a group buy, I'd recommend having a parts option as well, as I ended up ordering all the parts from 3 separate vendors. This option would be for those who don't really have a large stock of parts to tweak with, or just want to go with Peter's recommendations.
On another note, I recently ordered a custom R-Core 2x10 VAC 50VA transformer that I'll be able to compare to the Hammond EI that Peter recommends. It won't be here for a month or so.
Gopher said:
I'm not trying too be flattering here.......but Peter has alot to do with it. He has a proven track record when it comes to design and getting great results from minimal equipment. In a recent 6moons of Pass Labs XA30.5, Peter's Patek once again garnered some very positive comments when being compared to products two to four times it's cost.
As another illustration of this point. I have two NOS DACs. One is a TDA1541A and the other is Peter's TDA1543. For the cost Peter's TDA1543 does very well. I had to spend alot more money to better it.
Regards,
Dan
RC-EZ31 availability in Canada
First, a big thanks to Peter for sharing with us this
wonderful find and his modifications.
I just checked the bestbuy.ca site and it looks like the
RC-EZ31 is available again. You may be able to find it in you
local Best Buy stores as well. (The Toronto area stores
seem to have them.)
Dennis
First, a big thanks to Peter for sharing with us this
wonderful find and his modifications.
I just checked the bestbuy.ca site and it looks like the
RC-EZ31 is available again. You may be able to find it in you
local Best Buy stores as well. (The Toronto area stores
seem to have them.)
Dennis
Could someone please tell me - in layman's terms if possible - why it's important to set up the digital output to support a 75 ohm impedence, when it's accepted that a phono socket is not 75 ohms (even if the cable might well be)? Shouldn't it take into account the two phono connectors?
I guess the easy way out of this is to use a BNC? (which I shall experiment with when I get around to building a Shigaclone DAC).
Cheers,
- John
P.S. Just added (as I had it lying around) an Elna Silmic ARS 470uF / 63V to the C952 position, and it seems to have resulted in extra 'punch', and - interestingly - more delicacy on female vocals in particular. I know Peter left this position blank, but using the Elna seems to have resulted in an improvement.
- why it's important to set up the digital output to support a 75 ohm impedence, when it's accepted that a phono socket is not 75 ohms (even if the cable might well be)? Shouldn't it take into account the two phono connectors?I guess the easy way out of this is to use a BNC? (which I shall experiment with when I get around to building a Shigaclone DAC).
Cheers,
- John
P.S. Just added (as I had it lying around) an Elna Silmic ARS 470uF / 63V to the C952 position, and it seems to have resulted in extra 'punch', and - interestingly - more delicacy on female vocals in particular. I know Peter left this position blank, but using the Elna seems to have resulted in an improvement.
johnm said:Could someone please tell me - in layman's terms if possible
I guess the easy way out of this is to use a BNC? (which I shall experiment with when I get around to building a Shigaclone DAC).
Cheers,
- John
P.S. Just added (as I had it lying around) an Elna Silmic ARS 470uF / 63V to the C952 position, and it seems to have resulted in extra 'punch', and - interestingly - more delicacy on female vocals in particular. I know Peter left this position blank, but using the Elna seems to have resulted in an improvement.
Hi John,
C952 (470uF) in the RCEZ31 is the same location as C950 (1000uF) in the RCEZ32. I removed C950 which to my surprise resulted in quite a loss in performance. I then installed a 100uF Silmic and noted that it seemed much as before - perhaps even a bit better.
The board I am currently using has the stock for RCEZ32 1000uF cap. I will try a much better cap in this location when I have some time to listen critically again.
I have not yet started to mess with the tracking servo cap that everyone is wrestling with. I might leave it alone on this board.
On the spdif matching the idea is minimize impedance mismatches and the reflections they cause. Unfortunately RCA jacks and plugs do cause a mismatch - some are worse than others apparently, but without a TDR it is hard to tell. I have been evaluating the change with an actual 75 ohm bnc as opposed to the 50 ohm ones I've been using. Matching to the cable prevents signal fidelity issues that would otherwise occur - unmatched the cable looks like a lumped LC network which has an effect on bandwidth and group delay. I think it is pretty critical and I am trying to stick to the technically correct approach and figure out the interactions where possible and eliminate as many of them as I can. (FWIW I use all bnc on all of my homebrew digital gear, regrettably until recently I was having trouble sourcing 75 ohm hardware so some of it is still 50 ohm.)
Kevin I've just ordered a set of 4 75-ohm BNC connectors from Maplin. Already have some 75-ohm cable with connectors installed somewhere in my box of wires (THAT should take a day to find lol!).
I'll also replace the phono socket on my DPA DAC with the 75-ohm BNC connector (easy to do - just taken a look at it).
So as I'm technically installing a 'perfect' 75-ohm cable and connectors, which values do you think are the best to use at the SPDIF output? I'm currently using 291 / 100 ohms. Do you think I'd be better off with the 392 / 93.1 ohms combo?
Thanks,
- John
I'll also replace the phono socket on my DPA DAC with the 75-ohm BNC connector (easy to do - just taken a look at it).
So as I'm technically installing a 'perfect' 75-ohm cable and connectors, which values do you think are the best to use at the SPDIF output? I'm currently using 291 / 100 ohms. Do you think I'd be better off with the 392 / 93.1 ohms combo?
Thanks,
- John
johnm said:Could someone please tell me - in layman's terms if possible
Perhaps you should try have a look here
The pictures tell a lot about the problem. Every part has an effect on mismatching of the impedance. Connectors, cable, pcb tracks etc...
johnm said:
I've tried several combinations, but my primary concern is to match the coax and present a load to the chip that is within its safe output current capability. It may take you a while to find the combination that works best for you - I'm still working on that myself.
The 392/93.1 ohm combination is pretty standard for 5V cmos logic levels. 301/100 also gives you almost exactly 75 ohms - I've tried this combo and it seems to work without issues, output level is a bit hot, but I think within spec, and the current required of Dout is acceptable. (I'm currently using cheap 330 and 92 ohm 5% 1/2W CF resistors bought at Radio Shack - and I think I like these better than the 301/100 Holco combo, but I am not sure about that.
)
Thanks for the link Trilithium (Trek fan perchance?) - the pictures explain it well, though much of the txt went way over my head 
I guess the best we can do is ensure a steady 75 ohm output, transmission line, and receiver. I'll definately go with those 75 ohm BNC's. Maplin make a superb cable (for digital and analogue transmission) which I've been using for my digital connections recently.
"Manufacturer: Bandridge
• Double-shielded satellite cable
• Improved shielding for reduced interference
• 99.96% OFC conductor for crystal clear pictures and sound
• Air foam PE insulation layer for purest sound transmission
• Heavy-duty PVC jacket"
http://www.maplin.co.uk/module.aspx?ModuleNo=98026&doy=23m4
I use a long run of this cable to reduce the effect of any reflections when used as a digital interconnect. Definately the best cable I've used, and the cheapest!
Thanks for those figures Kevin. If it isn't too much trouble to write up, would you mind explaining how one comes up with those figures in the first place? I know the output of the 31 mech is -12mA. Min rating is listed as 0.9VDD. Erm..... how do I work this out?
- John
P.S. Couldn't resist - seeing as the Black Gates and Silmics seem to have burnt in sufficiently - I decided to have another go with that 0.1uF Auricap in the C906 position. MUCH more successful this time I'm pleased to say. Seems to have made the presentation more 'analogue' sounding.... more fluid and natural. This time the overall sound hasn't suffered. So please ignore what I said a few pages back folks - experiment with the 0.1uF cap, but ONLY after the caps have had a chance to burn in first.
I guess the best we can do is ensure a steady 75 ohm output, transmission line, and receiver. I'll definately go with those 75 ohm BNC's. Maplin make a superb cable (for digital and analogue transmission) which I've been using for my digital connections recently.
"Manufacturer: Bandridge
• Double-shielded satellite cable
• Improved shielding for reduced interference
• 99.96% OFC conductor for crystal clear pictures and sound
• Air foam PE insulation layer for purest sound transmission
• Heavy-duty PVC jacket"
http://www.maplin.co.uk/module.aspx?ModuleNo=98026&doy=23m4
I use a long run of this cable to reduce the effect of any reflections when used as a digital interconnect. Definately the best cable I've used, and the cheapest!
Thanks for those figures Kevin. If it isn't too much trouble to write up, would you mind explaining how one comes up with those figures in the first place? I know the output of the 31 mech is -12mA. Min rating is listed as 0.9VDD. Erm..... how do I work this out?
- John
P.S. Couldn't resist - seeing as the Black Gates and Silmics seem to have burnt in sufficiently - I decided to have another go with that 0.1uF Auricap in the C906 position. MUCH more successful this time I'm pleased to say. Seems to have made the presentation more 'analogue' sounding.... more fluid and natural. This time the overall sound hasn't suffered. So please ignore what I said a few pages back folks - experiment with the 0.1uF cap, but ONLY after the caps have had a chance to burn in first.
johnm said:Thanks for the link Trilithium (Trek fan perchance?) - the pictures explain it well, though much of the txt went way over my head
I guess the best we can do is ensure a steady 75 ohm output, transmission line, and receiver. I'll definately go with those 75 ohm BNC's. Maplin make a superb cable (for digital and analogue transmission) which I've been using for my digital connections recently.
"Manufacturer: Bandridge
• Double-shielded satellite cable
• Improved shielding for reduced interference
• 99.96% OFC conductor for crystal clear pictures and sound
• Air foam PE insulation layer for purest sound transmission
• Heavy-duty PVC jacket"
http://www.maplin.co.uk/module.aspx?ModuleNo=98026&doy=23m4
I use a long run of this cable to reduce the effect of any reflections when used as a digital interconnect. Definately the best cable I've used, and the cheapest!
Thanks for those figures Kevin. If it isn't too much trouble to write up, would you mind explaining how one comes up with those figures in the first place? I know the output of the 31 mech is -12mA. Min rating is listed as 0.9VDD. Erm..... how do I work this out?
- John
P.S. Couldn't resist - seeing as the Black Gates and Silmics seem to have burnt in sufficiently - I decided to have another go with that 0.1uF Auricap in the C906 position. MUCH more successful this time I'm pleased to say. Seems to have made the presentation more 'analogue' sounding.... more fluid and natural. This time the overall sound hasn't suffered. So please ignore what I said a few pages back folks - experiment with the 0.1uF cap, but ONLY after the caps have had a chance to burn in first.
Hi John,
I'll check and see if the Bandridge or equivalent cable is available here - is it flexible?
On selecting the resistors 3 criteria apply, first look at the minimum overall resistance values that do not call for more than 12mA, then look at the ratios that give you 75 ohm source impedance and result in output voltages that are within the allowable window (0.5Vpp when terminated 1Vpp when unterminated - note that I believe levels of 0.5Vpp - 1Vpp are considered acceptable when terminated. I've never seen spdif at levels significantly above 0.5Vpp in any commercial gear I've owned when terminated. Just math really...
I have more BG on the way and plan to wait to tweak that tracking servo cap until the rest of them are broken in..
Zen Mod said:if one can use sat cable as digital one , then one can use F connectors, too .
what's more logical .......
From an RF point of view, F-connectors are superior...
But if you want to be able to connect/disconnect the cable often,
you would destroy it in no time.
So BNC is the more logical choice.
Hi Kevin!
Thanks for the explanation.
The cable is fairly flexible - about the same flexibility as any old standard solid-core coax really.
Just been looking at my board, and the bridging cable on Okapi's PDF between the + pin of C916, to JW906. I'm having trouble tracing this on the circuit diagram - does this bypass C952 by any chance? (therfore making my 'upgrade' of an Elna Silmic 470uF in this position a waste of time?)
Thanks,
- John
Thanks for the explanation.
The cable is fairly flexible - about the same flexibility as any old standard solid-core coax really.
Just been looking at my board, and the bridging cable on Okapi's PDF between the + pin of C916, to JW906. I'm having trouble tracing this on the circuit diagram - does this bypass C952 by any chance? (therfore making my 'upgrade' of an Elna Silmic 470uF in this position a waste of time?)
Thanks,
- John