IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Is this RIAA EQ stage problem ?, ESI DuaFire's phono preamp gives too bright audio ...
Juha
post Aug 1 2010, 13:30
Post #1





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



Hi!

I have ESI DuaFire audio interface which has phono pre-amp onboard.
By the manufacturer, connection parameters are 47kΩ/200pF.
My cartridge specs says capacitance should be less than 200pF and the resistance 47kΩ.

Problem is that the resulting audio from DuaFire is (IMO) way too bright. I tried to mesure the RIAA curve using RMAA software but it didn't give fair results (lower curve (< 1kHz) looked 'bout well formed but the curve for upper frequencies (> 1kHz) didn't ... it never went under -5dB.
Maybe RMAA isn't right software for this type of measurements. Can anyone suggest proper software for this measurement?

Is it possible to localize the issue by comparing/examining the audio data?

http://www.hydrogenaudio.org/forums/index....showtopic=82617

Comparison is made against Terratec iVinyl USB which has connection parameters 47kΩ/100pF.

Juha


This post has been edited by Juha: Aug 1 2010, 13:34
Go to the top of the page
+Quote Post
Knowzy
post Aug 1 2010, 17:15
Post #2





Group: Members
Posts: 128
Joined: 15-July 08
Member No.: 55856



I'm not sure how you can use RMAA as it is intended to test a phono input. The samples you uploaded are music. RMAA is meant to analyze test tones.

The proper way to test the input is to feed in the test tones RMAA generates. You can do this using the output of your sound card.

The major roadblock is you outputs will be line level, not phono level.

The other problem is you may be measuring flaws in your output, which is why they suggest using a "reference sound card" to generate the test tones.

The software sounds very promising. Maybe others have ideas on setting up the conditions for proper testing.

FWIW, the TerraTec sounds more pleasing to my ears and jiiteepee at VinylEngine agrees with your analysis:

QUOTE (jiiteepee)
Resulting sound is (too) bright compared to what I've got from either Pro-Ject PhonoBox + E-MU 0404 USB or Terretec iVinyl USB.
Go to the top of the page
+Quote Post
Juha
post Aug 1 2010, 18:13
Post #3





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



QUOTE (Knowzy @ Aug 1 2010, 19:15) *
I'm not sure how you can use RMAA as it is intended to test a phono input. The samples you uploaded are music. RMAA is meant to analyze test tones.

The proper way to test the input is to feed in the test tones RMAA generates. You can do this using the output of your sound card.

The major roadblock is you outputs will be line level, not phono level.

The other problem is you may be measuring flaws in your output, which is why they suggest using a "reference sound card" to generate the test tones.

The software sounds very promising. Maybe others have ideas on setting up the conditions for proper testing.

FWIW, the TerraTec sounds more pleasing to my ears and jiiteepee at VinylEngine agrees with your analysis:

QUOTE (jiiteepee)
Resulting sound is (too) bright compared to what I've got from either Pro-Ject PhonoBox + E-MU 0404 USB or Terretec iVinyl USB.



Hmmm.. that was my post on VE (as well).

Actually, I tested as you mentioned. Plugged the DuaFire inputs into its outputs (loop back cabling) and made some measures. Was that right method for this type of measuring, I don't know.

I already downloaded Yoshimasa Electronic Inc's software which may be right software for this test - http://www.ymec.com
Tried two methods ... Software Out -> E-MU 0404 --> DuaFire Software In and just DuaFire alone but didn't get RIAA curve showing up on FR plot window. Need to test more tomorrow.


Juha
Go to the top of the page
+Quote Post
Knowzy
post Aug 1 2010, 19:04
Post #4





Group: Members
Posts: 128
Joined: 15-July 08
Member No.: 55856



QUOTE (Juha @ Aug 1 2010, 10:13) *
Hmmm.. that was my post on VE (as well).

I thought that quote was a little too perfect and timely! smile.gif

QUOTE (Juha @ Aug 1 2010, 10:13) *
Actually, I tested as you mentioned. Plugged the DuaFire inputs into its outputs (loop back cabling) and made some measures. Was that right method for this type of measuring? I don't know.

The problem is your line out is at line level. If you test with the input set to phono, you'll overload the input. If you test with the input set to line, you're not testing the pre-amp.

The trick is to get the test tones through the input at the proper level. You might be able to to this crudely through the headphone output with the volume set very low. There are probably other problems with using the headphone output, though I'm at a loss to say precisely what they are.
Go to the top of the page
+Quote Post
Knowzy
post Aug 1 2010, 19:46
Post #5





Group: Members
Posts: 128
Joined: 15-July 08
Member No.: 55856



One more thing occurred to me: The input gain is the other variable in the headphone out-to-phono in scenario.

To make sure things are consistent, leave the input gain at same level as when you record from your turntable and adjust your headphone volume only.
Go to the top of the page
+Quote Post
DVDdoug
post Aug 1 2010, 21:08
Post #6





Group: Members
Posts: 2596
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



I happen to have that song ripped to MP3 from CD, and the iVinyl version sounds more like the MP3 than the DualFire version.


I realize that you're trying to make some measurements to quantify the results, but from listening I agree with your conclusion that the DualFiree is too bright. I assume you also have an analog phono preamp for comparison?


I suspect it is an RIAA issue rather than a capacitance issue for a couple of reasons... Higher (parallel) capacitance tends to dull the high frequencies, not increase them*. And, I would interpret "less than" as "less than or equal to". After all, if 199pF is OK then 200pF should be OK too. And, the difference between the two recordings is NOT subtle! I'll bet there wouldn't be that much difference if you were to double the capacitance, or remove the capacitance entirely (which you can't, due to cable capacitance).


200pF, is ~40k ohms of capacitive reactance at 20kHz. So, that capacitor in parallel with a 47k resistor does lower the impedance and could affect the high frequency response depending on the source impedance of the cartridge.


To make measurements, you probably need an attenuator. I think phono preamps have gain of about 40dB at 1kHz, so a 40dB (100:1) attenuator would be about right. If you can't build one, a mic-to-line attenuator should work. Although, since there is about a 40dB difference between the low-frequency gain and the high-frequency gain, you might need 2 or more attenuators to make measurements in different bands. I suppose the "proper solution" would be to build a combination attenuator & inverse RIAA circuit.


* There could be some strange resonance between the capacitance and the cartridge's inductance that causes a high-frequency peak, but I don't think that ever happens.)

This post has been edited by DVDdoug: Aug 1 2010, 21:12
Go to the top of the page
+Quote Post
Juha
post Aug 2 2010, 04:52
Post #7





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



QUOTE
I assume you also have an analog phono preamp for comparison?


Yes.

QUOTE
I suspect it is an RIAA issue rather than a capacitance issue...


This is what I'm afraid of because of, ... it's harder to get fixed (I've opened the box so no warranty left).

QUOTE
To make measurements, you probably need an attenuator.
...
I suppose the "proper solution" would be to build a combination attenuator & inverse RIAA circuit.


If I apply inverse RIAA to a test signal file (sweep 20Hz-20kHz) and then measure the device using RMAA (or RA) ... resulting frequency responce should be 'bout straight line ... would that be suitable type of method (compared to what described above but done using plain software) ... I would feed the test signal through E-MU to the DuaFire phono inputs (I could also use additional mixer unit in between those two devices for to control signal levels if that's necessary).

Juha
Go to the top of the page
+Quote Post
cliveb
post Aug 2 2010, 12:08
Post #8


WaveRepair developer


Group: Developer
Posts: 842
Joined: 28-July 04
Member No.: 15845



QUOTE (DVDdoug @ Aug 1 2010, 21:08) *
I suspect it is an RIAA issue rather than a capacitance issue for a couple of reasons... Higher (parallel) capacitance tends to dull the high frequencies, not increase them*. And, I would interpret "less than" as "less than or equal to". After all, if 199pF is OK then 200pF should be OK too. And, the difference between the two recordings is NOT subtle! I'll bet there wouldn't be that much difference if you were to double the capacitance, or remove the capacitance entirely (which you can't, due to cable capacitance).

I agree with Doug that this is an EQ rather than capacitance issue. Listening to the two clips you posted, it sounds to me as if the DualFire clip simply has not had any RIAA EQ applied. Not only is it bright, there's no bass either - which is what you'd expect without RIAA EQ. (Fiddling with capacitance produces far more subtle effects, and only at the very top end of the frequency spectrum - bass is not affected). I'm not familiar with the DualFire - are you sure there isn't some switch or setting you have to adjust to put the RIAA EQ in circuit?
Go to the top of the page
+Quote Post
Juha
post Aug 2 2010, 16:19
Post #9





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



QUOTE
it sounds to me as if the DualFire clip simply has not had any RIAA EQ applied


This was my 1st thought as well after plugging the turntable and played something.

As I have made some RIAA EQ VST plug-ins, I tried by adding one into signal path in recording software end but, then audio was way too bassy and no treble at all so, the theory of fully missing RIAA EQ was turned down. Maybe it's just partly defected.

Specs paper the ESI support sent me didn't list the RIAA EQ accuracy nor gain value ... only input impedance/capacitance values.

QUOTE
are you sure there isn't some switch or setting you have to adjust to put the RIAA EQ in circuit?


I'm sure it's set properly. By switching to phono there's additional gain but I have not measured it yet.
While measuring the device using RMAA, difference in frequency plot between line and phono input was clear. As the loopback test method isn't suitable in this case, I got warned 'bout distortion > 6% and the frequency responce for frequensies > 1kHz didn't follow RIAA curve at all as I mentioned earlier (neither does frequencies < 1kHz).




Juha

This post has been edited by Juha: Aug 2 2010, 16:23
Go to the top of the page
+Quote Post
Juha
post Aug 2 2010, 18:19
Post #10





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



Damn .. the Edit button disappears too quickly.


Here's RIAA curve for Pro-Ject Phono Box MM got from RMAA (measured through path software->USB port->E-MU (Main Out) ->Pro-ject ->DuaFire (front Line-Ins) -> FireWire port -> software)




So, it looks like the RIAA pre-amp in ESI is broken ? Is it just the 2122Hz treble roll off missing ... (though, the lower end isn't well either)? If this is the issue then I need to prepare a software filter that handles missing eq.


Juha

This post has been edited by Juha: Aug 2 2010, 18:36
Go to the top of the page
+Quote Post
DVDdoug
post Aug 3 2010, 02:02
Post #11





Group: Members
Posts: 2596
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
So, it looks like the RIAA pre-amp in ESI is broken ?
Yes, but I'd use the word "defective". Since it's in both channels, that would tend to indicate a bad design, or more likely, the wrong filter components were installed during manufacture (wrong resistor or capacitor value). It's likely that they do some minimal "functional test" of the units coming off the assembly line, but they probably are not checking the EQ curve on every single unit.

Even though it's out of warranty it wouldn't hurt to send a note to the manufacturer. They may already know that they built a batch of bad units, and they might offer to repair it.

QUOTE
If this is the issue then I need to prepare a software filter that handles missing eq.
That's one solution, but since you have some other equipment I'd recommend using something else. Or, use the DuaFire's line inputs with your analog RIAA preamp.
Go to the top of the page
+Quote Post
Juha
post Aug 3 2010, 05:44
Post #12





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



QUOTE
QUOTE
If this is the issue then I need to prepare a software filter that handles missing eq.


That's one solution, but since you have some other equipment I'd recommend using something else. Or, use the DuaFire's line inputs with your analog RIAA preamp.


I already have much better audio interface for that purpose.

IIRC, regarding software based RIAA EQ filtering, someone mentioned once on another forum that the lowest area of RIAA EQ should be done in hardware to get the best out of software implementation. Maybe I'll just modify the DuaFire RIAA stage for that purpose.

Juha
Go to the top of the page
+Quote Post
WernerO
post Aug 3 2010, 07:39
Post #13





Group: Members
Posts: 74
Joined: 21-November 06
Member No.: 37858



Open it up again, take detailed pictures, trace the RIAA circuit (it won't be much) and then we can assist in repairing it...
Go to the top of the page
+Quote Post
Juha
post Aug 6 2010, 12:46
Post #14





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



QUOTE
Open it up again, take detailed pictures, trace the RIAA circuit (it won't be much) and then we can assist in repairing it...


Maybe this subforum isn't right place ...

Tracing the circuit is not an easy task for me but, hey, here are the pictures (I have only mobilephone camera (5mpix) in use so, the quality of takes isn't best possible):

Rear panel connections:

http://img39.imageshack.us/img39/5894/esiinpsw.jpg


Upper side of the board:

http://img826.imageshack.us/img826/7079/06082010566.jpg

Input port #1 is the upper rightmost connector in picture.


Other side of the board:

http://img251.imageshack.us/img251/5931/06082010570.jpg


Least values of most of those tiny components can be seen in those pictures.

Juha

This post has been edited by Juha: Aug 6 2010, 12:49
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Aug 16 2010, 12:55
Post #15





Group: Members
Posts: 3797
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (Juha @ Aug 6 2010, 07:46) *
QUOTE
Open it up again, take detailed pictures, trace the RIAA circuit (it won't be much) and then we can assist in repairing it...


Maybe this subforum isn't right place ...



I did lots of mods to audio gear back in the 60s, 70s, 80s and 90s. We had 1-sided circuit boards, wire leaded components, and even discrete semiconductors or simple ICs. Looking at your pcitures I see that 80s-style mods are probably impractical to apply to ca. 2010 equipment. SMT, multisided or even multi-layered boards, and highly integrated active components are game changers. Big time!

I think your best bet is to either use some audio interface or add outboard gear that works right as supplied, or see what you can do with software in the digital domain.
Go to the top of the page
+Quote Post
Juha
post Aug 16 2010, 17:09
Post #16





Group: Members
Posts: 457
Joined: 14-February 07
From: EU-FIN
Member No.: 40610



Yep, it's multi-layered board so it's not easy to fix if you don't have proper instruments for to do it.

On another forum, I got an suggestion to check the R1 from connection showen in picture:




Why, because if the resitance is too big, it's possible that the non-inverting opamp circuit hitts its minimum unity gain at far too low a frequency? I already simulated this in LTSpice and looks that the effect of too big resistance looks similiar showing in my picture of DuaFire measurement (post #9). If that's not the root for issue I'll try to fix the stage to do only the 3180μs pole.


Juha

This post has been edited by Juha: Aug 16 2010, 17:58
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 21st September 2014 - 05:22