I don't know anything about the cardas disc. What I do know is that the test records from CBS are great. Sometimes they are available on Discogs. I got lucky and imported a factory sealed CBS 130 record from the USA. I shared my test results with Room EQ Wizard on the vinyl engine forums.
I have no idea why anyone would go through all the trouble to record a test LP just to have the frequency sweep begin at 1kHz...
There was a problem in the early firmware releases that they hadn't really considered signals below 20Hz. There is a special Subwoofer notch function but I have not found anything so say how a (say) 30Hz HPF would actually work with current firmware.
Good question and we are really not sure. When we tried to compare test records there were some anomalies. Avoid pink noise tracks as they vary considerably.
What cartridge and why do you think you need to adjust RIAA for the cartridge? (honest question)
For me it's AT132 w/ ATN152. I EQ'd it to be flat after RIAA.
Before I did this I heard clear differences in tonality between LPs and CDs with the same content.
In a system that is EQ'd flat and has FIR room correction, the errors in tonality stick out like a sore thumb.
I use the ADAU DSP devices in a custom coded system and have no problems with low frequency filters. I use subsonic cutoffs at 10 and 15 Hz regularly.
You might want to check the busy at coefficients minidsp is loading if you can.
On the pop from your turntable... is the mute relay creating an open or short on the input to your preamp? If so there is every chance you have a DC offset say from one input bias current to the op amp that is being altered--> pop.
You might want to check the busy at coefficients minidsp is loading if you can.
On the pop from your turntable... is the mute relay creating an open or short on the input to your preamp? If so there is every chance you have a DC offset say from one input bias current to the op amp that is being altered--> pop.
47K and about 100pf of cable. Any other combination is less linear than that.
I plan to split the signal into mid and side, then use an LR8 highpass at 18Hz on the mid signal. Great to know that this will probably work.
The mute relay shorts the turntable output when active.
The circuit I use is basically this:
VSPS
The changes are R2 = 220 Ohm, R4 = 22K, R3, C2 and C3 have been removed.
I use a symmetrical power supply. If I understand correctly, that means there should not be any DC bias current to go into the opamp. Is that right? I'm not trying to be smart here, I genuinely do not know. All help is very much appreciated.
DL103 (regular one). It came with FR paper, but I'm not sure how accurate it is, and it could be an ornamental thingy for audiophile.
hahaha, it sounds like it would make things more confusing than having no reference sweep record. 😱
Verification>?
Have you actually measured the response of those low frequency digital filters to verify they perfectly match the roll off frequency and slope you are designing to. Digital filters hit a wall due to limited precision of the 28 bit word and 56 bit accumulator when running 96 kHz sample rate and low frequency filters. Maybe you are running lower sample rate for the low frequency filters or double precision coefficients ( 56 bits ) or something. You can see the effect in simulation as well as in the firmware. In the miniDSP 2x4hd the filters stop matching the prototype curves in the 30 Hz range. Just a limitation of the bit depth and sample rate whether you use the built in filter tool or load coefficients you compute yourself. I would like to see miniDSP add a sample rate conversion and run the low frequency filters at a lower sample rate, say 11 ksps, so they would work properly with the limited precision math.
Have you actually measured the response of those low frequency digital filters to verify they perfectly match the roll off frequency and slope you are designing to. Digital filters hit a wall due to limited precision of the 28 bit word and 56 bit accumulator when running 96 kHz sample rate and low frequency filters. Maybe you are running lower sample rate for the low frequency filters or double precision coefficients ( 56 bits ) or something. You can see the effect in simulation as well as in the firmware. In the miniDSP 2x4hd the filters stop matching the prototype curves in the 30 Hz range. Just a limitation of the bit depth and sample rate whether you use the built in filter tool or load coefficients you compute yourself. I would like to see miniDSP add a sample rate conversion and run the low frequency filters at a lower sample rate, say 11 ksps, so they would work properly with the limited precision math.
There is also a 1 kHz to 20 Hz sweep. Splitting the sweeps at 1K was done on many test LPs to improve S/N ratios.
CBS STR 130 This is the one I have used successfully. There was a user on the vinylengine forum who had the STR 100 record and had successfully used it to get a frequency response graph using ARTA.
I missed that, you're right.
I also have this one:
Elipson Test Record
It looks like it should work just as well but I have not tested it yet. As the sweep starts at 25Hz it might even be more useful than the CBS STR 130. I think the next time I measure I will use this one.
In the mean time I got impatient and hacked up a cable to connect my Sure DSP to the opamp board I made. The RIAA biquads work great!
It's just my board... I probably made every layout mistake known to man. As soon as the needle hits the records and my player unmutes there is a significant amount of hiss. When the relay shorts the player's outputs the hiss is gone - the problem is obviously in my circuit. Also I have a huge difference in gain between both channels. One channel - the one that is more quiet - sometimes makes noises akin to tuning a shortwave receiver. I guess I need to start from scratch on this board and salvage what I can (my soldering is not exactly great either).
Well at least there is progress 🙂
The CBS test LP's are full of problems (I have all of them BTW). An unequalized constant velocity sweep has to start at 500-1000Hz for obvious reasons. Old threads don't give immunity to misconceptions, a bad sign.
Last edited:
Is there one you can recommend that is better than those by CBS? I have the Elipson I mentioned before and the QR2010 by Brüel & Kjær to try out.
The only totally clean pink noise was on the CH Labs 45 RPM disk that came with their $$$ phono pre-amp. The bad tracks can be easily identified by playing them off speed, the badness is on the disk not your equipment. The best sweep on the CBS LP's is the 500Hz - 50kHz constant velocity track (IIRC STR140), no equalization circuitry in the way. You have to play this with a flat pre-amp these are not consumer items.
Last edited:
I found every pink noise track I have used to be lacking as well. I made the frequency response graph I made of my setup with the logarithmic sine sweep from CBS STR 130.
Is there a reason why I should use pink noise? The only usable results I got so far were from long sine sweeps.
Edit: Just noticed your edit, thanks.
Last edited:
Looking at that plot of yours, and based on what Hans found out about AT generators I would try upping the Resistive load to 65k for 100pF. But I think you might still have more capacitance than you think lurking somewhere.
Each to their own but I like to flatten the generator response as much as possible before correcting any residual.
Once again, flattening the curve seems like the right thing to do.
I'll experiment some more as soon as I have my opamp mess sorted.