advertisement


Leaving Paradise -new phono stage

Arkless Electronics said:
I found Paradise RIAA to be very accurate. Certainly within 0.1-0.2dB limits but I also worked out values to correct even that to theoretical 0.01dB type accuracy... not that I reckon that's going to make any difference!
Click to expand...
Interesting as mastermind behind Paradise concept says that he took funky approach to RIAA network and made it to sound good with lesser recordings by rising bass and treble , respectively by 0.6 and 0.4dB.
I tested this values and I’ve had similar results , maybe 0.1-0.2dB out , so did few other guys on DIYAudio ,Spice simulation confirmed this .
How do you test RIAA network anyway, I assume you do spot check at chosen frequencies with signal generator?
 
slawekt said:
Interesting as mastermind behind Paradise concept says that he took funky approach to RIAA network and made it to sound good with lesser recordings by rising bass and treble , respectively by 0.6 and 0.4dB.
I tested this values and I’ve had similar results , maybe 0.1-0.2dB out , so did few other guys on DIYAudio ,Spice simulation confirmed this .
How do you test RIAA network anyway, I assume you do spot check at chosen frequencies with signal generator?
Click to expand...

Both SPICE simulation and I have reverse RIAA generator and Bruel & Kjaer signal generator plus widerange AC millivoltmeter.

I only simulated it for the Paradise but it can be done more accurately in simulation than in real life as a perfect reverse RIAA generator can be used virtually which is built with +/- 0.00000% accurate parts. This can then allow you to design RIAA networks of extreme accuracy which will be limited only by the tolerance of the real components.

When I design and build phono stages I use the reverse RIAA generator to check the actual response as a "reality check". With many designs the topology is such that if the simulation says the RIAA is spot on then it will be in reality but it can be useful to check that eg the 1% parts used really live up to that spec and that nothing has drifted.

In some designs the accuracy of the RIAA can be affected by the topology and the parts used, I'm mainly thinking of various simple valve phono stages here in which the sample of a particular valve used and how worn it is can all effect the RIAA accuracy.... Not all valve phono stages are affected by this but many of the simpler and more popular ones are.

As you may be aware I designed and built a phono stage working on the same virtually unique transconductance principle as The Paradise from ground up. It uses completely different topology and was intended specifically to take on The Paradise..... Which it very successfully managed to do! (Only 1 exists and they are not available to buy so no enquiries please folks).
 
Do you know what RIAA network has been used in the one you’ve tested ,

I checked wide range of frequencies from laboratory signal generator without inverse RIAA also done frequency sweep and white noise response and standard RIAA isn’t flat down to 0.1-0.2 dB in my experience?

What’s the RIAA node impedance when simulated in Spice I’m curious?
 
slawekt said:
Do you know what RIAA network has been used in the one you’ve tested ,

I checked wide range of frequencies from laboratory signal generator without inverse RIAA also done frequency sweep and white noise response and standard RIAA isn’t flat down to 0.1-0.2 dB in my experience?

What’s the RIAA node impedance when simulated in Spice I’m curious?
Click to expand...

It was in a sim as I outlined above and was the "standard" RIAA as published in the schematic for revision 3.... and erm.... I was wrong and you are right... I somehow had it in my memory that the RIAA was super accurate but I've just opened it in my simulator and had another look and it is inaccurate exactly as you describe. +0.45dB at LF and +0.4dB at HF in my sim at final shelving points. This will be why I designed a new and accurate RIAA network for The Paradise:rolleyes:

Do you really need that node impedance? I may get round to it later if it matters to you:)
 
I've found values to get it within 0.02dB if anyone wants to try it.

Change: 73.5K to 72K, 9K1 to 9K8, 33.3nF to 32.7nF and 11nF to 11.1nF. Leave the 220R shorted or remove and place wire link in. I believe it's generally shorted by most users and if not then it causes a further kick up in HF above about 18KHz....
 
Arkless Electronics said:
It was in a sim as I outlined above and was the "standard" RIAA as published in the schematic for revision 3.... and erm.... I was wrong and you are right... I somehow had it in my memory that the RIAA was super accurate but I've just opened it in my simulator and had another look and it is inaccurate exactly as you describe. +0.45dB at LF and +0.4dB at HF in my sim at final shelving points. This will be why I designed a new and accurate RIAA network for The Paradise:rolleyes:

Do you really need that node impedance? I may get round to it later if it matters to you:)
Click to expand...
Don't worry about ,I was a bit concerned that my measurements and simulations were so far away from yours
 
Arkless Electronics said:
I've found values to get it within 0.02dB if anyone wants to try it.

Change: 73.5K to 72K, 9K1 to 9K8, 33.3nF to 32.7nF and 11nF to 11.1nF. Leave the 220R shorted or remove and place wire link in. I believe it's generally shorted by most users and if not then it causes a further kick up in HF above about 18KHz....
Click to expand...

In practice RIAA node impedance is higher than simulated and 72k has to be dropped down a bit but your values are very close , anyone else tested RIAA accuracy of their Paradise?
 
You all know this but boosting bass and treble (+0.4db as mentioned above) is equivalent to having what we used to call a "loudness contour" to compensate for the fall off in human hearing as the sound level is reduced (Fletcher-Munsen curve). It seems that the Paradise designer wanted his unit to sound flat at a particular playback level - not necessarily full volume. Perhaps the Paradise should have a switch (flat or not) so people can do direct comparisons at the same volume level? Otherwise, the louder unit will almost always sound better to the listener. One of those testing quandaries.
 
The loudness contour as used on a button is typically about 4db. 0.4db is barely audible.

And one would hope any testing would be level matched to the gain of each phonostage.

But yes, the Paradise eq is designed to compensate for a flatter response at a specific listening level.
 
  • Like
Reactions: GT
Arkless Electronics said:
.... . The Paradise..... Which it very successfully managed to do! (Only 1 exists and they are not available to buy so no enquiries please folks).
Click to expand...


Jez, it's getting a bit tiresome hearing of your advanced design skills and you not having a schematic or product to prove it, please produce a schematic, product or keep quiet!
 
russel said:
Jez, it's getting a bit tiresome hearing of your advanced design skills and you not having a schematic or product to prove it, please produce a schematic, product or keep quiet!
Click to expand...

Never in a million years pal! And yes I'm one of the best designers in the world thank you:p
 
sq225917 said:
The loudness contour as used on a button is typically about 4db. 0.4db is barely audible.

And one would hope any testing would be level matched to the gain of each phonostage.

But yes, the Paradise eq is designed to compensate for a flatter response at a specific listening level.
Click to expand...

Nah it has to be accurate AFAIC... anything else is "cheating" and and unfaithful to the tenets of hi fi;)
 
YNWOAN said:
Frankly, if you just want to try something different you should change every other part of your system first. I’ve heard most of the options open to you and they are nowhere near what you have. Having said that, it’s more than likely that the rest of your system is actually holding it back - sounds harsh but is likely true.

Edit: lots of room for error in this build though - so it all depends....
Click to expand...
You're correct there is a lot of room for error , I hope your Paradise is not the one with LED's taken out of regulator Vreference branch and fitted to illuminate the casing
Anyway , after some busy at work time I will be fitting some custom made Vishay naked foil resistors into RIAA network , regulator needs capacitor upgrade (as found out by one of DIYAudio guys) , some silver input and output wiring , final adjustment and another , probably last Paradise-Vida shoot-out .
I have lived with Vida for all this time and to be honest I haven't missed Paradise but let's see , hopefully all bits will arrive before my holiday .
 
No, my LED’s are all in their correct places. Mine has a smoked acrylic case and you can see when it’s on (it’s also got all the bells and whistles you mention).

I was kindly loaned a Vida and thought it was excellent - I could easily have lived with it, and happily so too.
 
You might need to check C103,C203 capacitors in the regulator as they seem to be under rated , you need something like 25V rating . After reading Paradise thread I have pulled mine out and they give bit random readings far bit above 100uF .
I'd also advice to make some probing with a good oscilloscope (80-100Mhz) , voltmeter is really not enough with circuit like this .
If anyone is interested I can check your boards .
 
Many people built their paradise with the leds lifted 1cm or so from the boards, to make it easier to desolder/replace them, none reported any issues. It's the same total length as lifting one to the front panel. They measure just the same, it's a none issue. Rigol 70mhz scope.
 
If someone who designed the circuit says to avoid this I'd rather listen to him , regulator works on the border when it comes to HF stability and these LED's provide reference for the whole amplifier .Why to induce another potential source of instability while circuit is barely stable thanks to the compensation caps also as Salas says extended wire works like an antenna .
Tektronix 100MHz scope
 
You must log in or register to reply here.


advertisement


Back
Top