• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

ancient history: M-Audio Audiophile 2496 and Delta 410 cards from close to 25yrs ago

Barrelhouse Solly

Senior Member
Joined
Aug 13, 2020
Messages
375
Likes
359
The M-Audio 2496 was my first fancy sound card. I bought it to do home recording and made demos and youtube slide shows with it. I bought my first decent quality mic to use with it. I have fond memories. I replaced it nearly 20 years ago with a 2 input Focusrite interface.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,670
Likes
38,766
Location
Gold Coast, Queensland, Australia
Why is it 24.576 MHz? I would expect 28.224 MHz, since that is the least common multiple of the 6 most common sample rates (44.1, 48, 88.2, 96, 176.4 and 192).

It's the standard for audio on cards/firewire/DAT etc- (512x48kHz) and because that is the crystal on the board...

1693435943462.jpeg


I reckon a 24.576 MHz TCXO (3.3/5V) and two identical short lengths of coax to the clock pins (cut the tracks) would synchronize them from a hardware perspective.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,476
Likes
4,094
Location
Pacific Northwest
It's the standard for audio on cards/firewire/DAT etc- (512x48kHz) and because that is the crystal on the board...
...
Since 24.576 is an integer multiple of 48 / 96 / 192, but not of 44.1 / 88.2 / 176.4, I wonder if this board has slightly worse measurements at those latter rates.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,371
Likes
18,279
Location
Netherlands
Since 24.576 is an integer multiple of 48 / 96 / 192, but not of 44.1 / 88.2 / 176.4, I wonder if this board has slightly worse measurements at those latter rates.
It does:

What the results show, is that the card is apparently capable of the specified dynamic range (116 dB) and THD+N (-105 dB) at a sample rate that is a multiple of 48 kHz, but the noise performance degrades significantly at 44.1/88.2/176.4 kHz.
 

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Why is it 24.576 MHz? I would expect 28.224 MHz, since that is the least common multiple of the 6 most common sample rates (44.1, 48, 88.2, 96, 176.4 and 192).
The 24 denotes the amount of bits and the 96 the hz, so the card has a maximum of 24 bits at up to 96khz.
 

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Got one, have to get a PCI to PCIE but really excited to give this a go.

maudio 2496.jpg


Got some other new toys that should give me a reliable reference.

photo_1_2023-09-07_14-01-30.jpg


photo_2_2023-09-07_14-01-30.jpg



Got to get me a Dayton imm-6 or the UMM-6 a start to take some measurements.
 

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,476
Likes
4,094
Location
Pacific Northwest
Why is it 24.576 MHz? I would expect 28.224 MHz, since that is the least common multiple of the 6 most common sample rates (44.1, 48, 88.2, 96, 176.4 and 192).

The 24 denotes the amount of bits and the 96 the hz, so the card has a maximum of 24 bits at up to 96khz.
Sure, but that's not what I was asking. We were talking about the board's master clock frequency, not the bit depth or sample rate of the music.
 

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Sure, but that's not what I was asking. We were talking about the board's master clock frequency, not the bit depth or sample rate of the music.
Sorry it is me again looking at a single line of text and making my own interpretations instead of looking at the whole, my cart is very often before my horse.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,373
Likes
3,318
Location
.de
Got one, have to get a PCI to PCIE but really excited to give this a go.
Don't expect too much by modern standards. Analog fullscale level is about 1.4 Vrms if memory serves (so short of 2 Vrms; spec is +2 dBV in/out), and the unbalanced I/O is going to invite ground loops if you are not super careful. You should still be able to accommodate the outputs with custom cabling if the pro gear visible in the photo is to be used (you are likely to need/want a mixer / monitor controller as well). The input side I'd only use with a phonopre, a portable mixer with an external power supply or perhaps the tape loop of a consumer integrated amp, with zero additional earth connections involved.

This was decidedly a "prosumer" card from a transitional period. People liked to use it for vinyl rips and things like that (the ADC digital filter is approaching pro-grade and a major cut above consumer offerings of the time and should be OK-ish at 44.1 kHz and fine at 48 kHz up, distortion is low, and the fixed-level line input generally provides enough headroom unless you're unlucky). It was a major step up for many at a time when consumer soundcards were quite frankly pretty lame and you could consider yourself lucky if yours sported an AK4531 or one of the Crystal AC97 codecs rather than say the Sigmatel jobs that SB Live! cards were rocking, and 87 dB(A) of dynamic range in loopback was considered good. (Don't even ask about the filter performance.) The AP2496 handily added 10 dB!
(Yes, pro/prosumer soundcards had existed for years at this point but remained decidedly niche.)

The more expensive successor Audiophile 192 then offered actual balanced I/O and good midrange converters (ADC AK5385A, DAC AK4358), but I don't think it ever was quite as popular. That was the time when the E-MU cards started to rock the world, and a 1212M wasn't that much more expensive and sported first-rate converters (AK5394 and CS4398 - almost 117 dB(A) in loopback and super low THD).
 

wwenze

Major Contributor
Joined
May 22, 2018
Messages
1,310
Likes
1,871
Most computer sound cards have a crystal meant for 48kHz, I can't remember all the long history already, but hey DVD is also 48kHz, AC97 is 48kHz...

To output an incoming 44.1kHz signal meant internal resampling to 48kHz output, except for a selected few of cards. Cards running on the Envy chipset and have 2 crystals like that M-Audio there would be able to use both frequencies with clock crystal. Even then there are limitations, like if you're not running in "exclusive mode" (or its historical equivalent) which didn't really exist back then anyway you'd be relying on either Windows or the driver to auto-switch (yea it's a feature of the driver to select which crystal) and most of the time it practically cannot auto-switch because the OS mixer is constantly running and it needs to make sure it can output concurrent streams of different source sampling rates so end up it's stuck at one frequency anyway, not counting going into settings and changing soundcard sampling rate which every card could do (at least on the surface, but inside? See earlier in paragraph)

But manual-switching is allowed and the clock control feature of the Envy chip can perform some funky behavior. Like you can change the clock on the fly and see your music go slower and pitch lower, which is a good way to teach people about master clock and "pulling" data. Yet do the same thing when playing a video game and noting how the sound still comes out at the some timings when you shoot a gun (albeit with a lower pitch) and you can teach about "pushed" data. You can even select external source (SPDIF) for your clock and then disconnect the SPDIF cable to stop the music entirely, then plug it back in and it resumes from the same position. (I presume different programs may have different behavior, see the music vs video game example) And lastly, you can try recording from SPDIF source but choose the internal clock and hear the resulting recording make noise at regular intervals, which teaches that no two crystals are the same frequency which is why we need either a shared master clock or clock recovery from the incoming signal, or some form of ASRC.
 

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,066
Likes
10,903
Location
São Paulo, Brazil
M-Audio Audiophile 24/96 was my first audiophile source, and later the ESI Juli@. Blast from the past indeed!
 

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Don't expect too much by modern standards. Analog fullscale level is about 1.4 Vrms if memory serves (so short of 2 Vrms; spec is +2 dBV in/out), and the unbalanced I/O is going to invite ground loops if you are not super careful. You should still be able to accommodate the outputs with custom cabling if the pro gear visible in the photo is to be used (you are likely to need/want a mixer / monitor controller as well). The input side I'd only use with a phonopre, a portable mixer with an external power supply or perhaps the tape loop of a consumer integrated amp, with zero additional earth connections involved.

This was decidedly a "prosumer" card from a transitional period. People liked to use it for vinyl rips and things like that (the ADC digital filter is approaching pro-grade and a major cut above consumer offerings of the time and should be OK-ish at 44.1 kHz and fine at 48 kHz up, distortion is low, and the fixed-level line input generally provides enough headroom unless you're unlucky). It was a major step up for many at a time when consumer soundcards were quite frankly pretty lame and you could consider yourself lucky if yours sported an AK4531 or one of the Crystal AC97 codecs rather than say the Sigmatel jobs that SB Live! cards were rocking, and 87 dB(A) of dynamic range in loopback was considered good. (Don't even ask about the filter performance.) The AP2496 handily added 10 dB!
(Yes, pro/prosumer soundcards had existed for years at this point but remained decidedly niche.)

The more expensive successor Audiophile 192 then offered actual balanced I/O and good midrange converters (ADC AK5385A, DAC AK4358), but I don't think it ever was quite as popular. That was the time when the E-MU cards started to rock the world, and a 1212M wasn't that much more expensive and sported first-rate converters (AK5394 and CS4398 - almost 117 dB(A) in loopback and super low THD).
I remember getting the e-mu 1212m card with the breakout box and sent it back because it did not give me the volume of the the 24/96, this was the best thing to happen as I then got the mbox1/digi002/mbox mini and the mbox 2 that I still have and is pictured above, my last place had good grounding, the next required me to lift the ground pin on the monitor(screen) and the place I am in now, some analog amps such as the samson servo 260 would give hideous noise from mouse movements until i lifted the ground pin on the pc's power supply, had to pull every damn plug in here to get to that conclusion.

Some pics of the card and it's chips.

photo_1_2023-09-08_01-16-14.jpg
photo_2_2023-09-08_01-16-14.jpg
photo_3_2023-09-08_01-16-14.jpg
photo_4_2023-09-08_01-16-14.jpg
photo_5_2023-09-08_01-16-14.jpg
photo_6_2023-09-08_01-16-14.jpg
photo_7_2023-09-08_01-16-14.jpg
photo_8_2023-09-08_01-16-14.jpg
photo_9_2023-09-08_01-16-14.jpg
photo_10_2023-09-08_01-16-14.jpg



Will try the unbalanced analogue outs but will have to use the spdif to go into a converter to optical for my AVI's, so might try that for the unbalanced analog outs too, no desk or pre amps here.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,373
Likes
3,318
Location
.de
my last place had good grounding, the next required me to lift the ground pin on the monitor(screen) and the place I am in now, some analog amps such as the samson servo 260 would give hideous noise from mouse movements until i lifted the ground pin on the pc's power supply, had to pull every damn plug in here to get to that conclusion.
Unsafe bodgery is always an option, I guess. You know, just because something "works" doesn't mean it's a good idea.

A PC is an IEC Class I device. The premise behind the construction regarding electrical safety is that if for some reason the mains live ever were to come in contact with chassis internally, it would find itself shorted to ground via the (ominously named) protective earth conductor, tripping a breaker or at least blowing the internal fuse in short order. This requires a permanent connection capable of carrying a substantial fault current on the order of at least 15 amps in 115V territory. Does audio wiring qualify as such? No? I didn't think so either.

This is why you do not tape off or insulate the mains earth connection on an IEC Class I device, period. Maybe briefly for troubleshooting, but sure as hell not long-term. I know "cheater plugs" used to be somewhat popular at a time, but just because you can buy something doesn't mean it's safe (you bet they're not UL listed).

Now would be a good time fo you to get up to speed regarding how to deal with ground loop issues properly. I know it's shocking, but the information is available on this very forum (gasp!), including how to make custom cables to deal with gear afflicted by the Pin 1 Problem and adapter cables for unbalanced outputs into balanced inputs (though whether I'll be able to find all of that in a hurry is another matter). Honestly I'm so over ground loops at this point, this is really getting old.


I hope you're not using the unbalanced RCA inputs on the Servo 260 when a PC is the source and there is no galvanic isolation of any kind in between, 'cause that's hopeless. The balanced inputs seem well-constructed (clearly these folks knew what they were doing even pre-AES48), but you obviously have to make use of it and not screw it up using improper cabling (*cough* instrument cable *cough*).
servo260-inputs.png

At the very least, you may consider investing the princely sum of 29 USD into a Behringer HD400 (outrageous, I know). Or perhaps even more than one, I don't know how complex your setup is. At the levels that typically go into a speaker amp (~1 Vrms give or take), it's pretty much transparent, even if it should be driven by a source that doesn't balk at the idea of driving a 600 ohm load and has as low an output impedance as possible, and cable length on the output side should be kept fairly low like no more than 5' (1.5 m) or so (none of which should surprise anybody who knows anything about the characteristics of transformers). Been using one effectively as an unbalanced to balanced converter for my monitors for years. It used to be more effective with my old Xonar D1 than my current onboard audio (I can hear low-level clicks whenever my fridge turns on and off now, though that's it), but still.

I had ordered a Xonar SE with the new computer mostly out of curiosity, just to see how a soundcard based on an ALC1220 does, but it proved too much of a hassle and not worth the extra power usage in the long term. (I was very proud of having built a machine that would idle at slightly less than 15 watts after some tweaking.) The card's USB controller is bug central, and it doesn't get along with ASPM. So onboard ALC1200 it is. Technically a downgrade, but good enough for the mundane task of driving active speakers to modest levels.
Will try the unbalanced analogue outs but will have to use the spdif to go into a converter to optical for my AVI's, so might try that for the unbalanced analog outs too, no desk or pre amps here.
Optical is good, that can keep you out of a lot of trouble. Now, for obvious reasons, the AP2496 shouldn't sound any different than any other digital source capable of 24-bit output under these circumstances (even lowly onboard audio), but hey. You shouldn't have any issues with jitter at least, with the Envy24 reigning over two different clock crystals for both 44.1 and 48 kHz family sample rates.

I used to rock two Terratec Aureon Space 5.1 cards (Envy24HT with WM8770 codec) for years, one of them reflashed to the very similar Audiotrak Prodify 7.1 so the Windows drivers would coexist. One for general playback, the other for radio recording. Had to mess with PCI latency so that the USB 1.1 controller on my BX board would not sabotage the audio. (USB 2.0 should be more cooperative, they switched to using busmastering then!) The Envy24 family of chips were great for low latency but easy to upset as well. One of the cards started suffering from dropouts towards the end, I think one of the (small and unremarkable quality) electrolytics was starting to go after a fair bit of use...
 
Last edited:

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Unsafe bodgery is always an option, I guess. You know, just because something "works" doesn't mean it's a good idea.

A PC is an IEC Class I device. The premise behind the construction regarding electrical safety is that if for some reason the mains live ever were to come in contact with chassis internally, it would find itself shorted to ground via the (ominously named) protective earth conductor, tripping a breaker or at least blowing the internal fuse in short order. This requires a permanent connection capable of carrying a substantial fault current on the order of at least 15 amps in 115V territory. Does audio wiring qualify as such? No? I didn't think so either.

This is why you do not tape off or insulate the mains earth connection on an IEC Class I device, period. Maybe briefly for troubleshooting, but sure as hell not long-term. I know "cheater plugs" used to be somewhat popular at a time, but just because you can buy something doesn't mean it's safe (you bet they're not UL listed).

Now would be a good time fo you to get up to speed regarding how to deal with ground loop issues properly. I know it's shocking, but the information is available on this very forum (gasp!), including how to make custom cables to deal with gear afflicted by the Pin 1 Problem and adapter cables for unbalanced outputs into balanced inputs (though whether I'll be able to find all of that in a hurry is another matter). Honestly I'm so over ground loops at this point, this is really getting old.


I hope you're not using the unbalanced RCA inputs on the Servo 260 when a PC is the source and there is no galvanic isolation of any kind in between, 'cause that's hopeless. The balanced inputs seem well-constructed (clearly these folks knew what they were doing even pre-AES48), but you obviously have to make use of it and not screw it up using improper cabling (*cough* instrument cable *cough*).
View attachment 310612
At the very least, you may consider investing the princely sum of 29 USD into a Behringer HD400 (outrageous, I know). Or perhaps even more than one, I don't know how complex your setup is. At the levels that typically go into a speaker amp (~1 Vrms give or take), it's pretty much transparent, even if it should be driven by a source that doesn't balk at the idea of driving a 600 ohm load and has as low an output impedance as possible, and cable length on the output side should be kept fairly low like no more than 5' (1.5 m) or so (none of which should surprise anybody who knows anything about the characteristics of transformers). Been using one effectively as an unbalanced to balanced converter for my monitors for years. It used to be more effective with my old Xonar D1 than my current onboard audio (I can hear low-level clicks whenever my fridge turns on and off now, though that's it), but still.

I had ordered a Xonar SE with the new computer mostly out of curiosity, just to see how a soundcard based on an ALC1220 does, but it proved too much of a hassle and not worth the extra power usage in the long term. (I was very proud of having built a machine that would idle at slightly less than 15 watts after some tweaking.) The card's USB controller is bug central, and it doesn't get along with ASPM. So onboard ALC1200 it is. Technically a downgrade, but good enough for the mundane task of driving speakers.

Optical is good, that can keep you out of a lot of trouble. Now, for obvious reasons, the AP2496 shouldn't sound any different than any other digital source capable of 24-bit output under these circumstances (even lowly onboard audio), but hey. You shouldn't have any issues with jitter at least, with the Envy24 reigning over two different clock crystals for both 44.1 and 48 kHz family sample rates.

I used to rock two Terratec Aureon Space 5.1 cards (Envy24HT with WM8770 codec) for years, one of them reflashed to the very similar Audiotrak Prodify 7.1 so the Windows drivers would coexist. One for general playback, the other for radio recording. Had to mess with PCI latency so that the USB 1.1 controller on my BX board would not sabotage the audio. (USB 2.0 should be more cooperative, that used busmastering!) The Envy24 family of chips were great for low latency but easy to upset as well. One of the cards started suffering from dropouts towards the end, I think one of the (small and unremarkable quality) electrolytics was starting to go after a fair bit of use...
Thanks, that is some very detailed info and great tips on the ground issue, I forgot to state that the pin was only lifted for a short while to identify the cause and a new power supply purchased which has seemed to have dome the trick.
 

Citizen

Member
Joined
Apr 23, 2023
Messages
63
Likes
36
Thanks, that is some very detailed info and great tips on the ground issue, I forgot to state that the pin was only lifted for a short while to identify the cause and a new power supply purchased which has seemed to have dome the trick.
A Behringer HD400 is on it's way to me.

Again thanks.
 
OP
T

TonyJZX

Major Contributor
Joined
Aug 20, 2021
Messages
1,974
Likes
1,920
strangely enough both of my main pcs have pci so i could fish out the cards and see if they work

but:

M-Audio Delta 410 Drivers Download​


  • Driver Version: 5.10.0.5078
  • Release Date: 2011-12-21
  • File Size: 109.22K
  • Supported OS: Windows 10 64 bit, Windows 8.1 64bit, Windows 7 64bit, Windows Vista 64bit, Windows XP 64bit
released date?
 

audio_tony

Addicted to Fun and Learning
Joined
Feb 24, 2019
Messages
574
Likes
697
Location
Leeds, UK
I disagree with those Head-Fi* measurements of the Asus Xonar DX card - I have the same card and I get excellent measurements out of it.

96k / 4kHz tone - 0.00034% THD - noise floor approaching -160dB. This is a standard card with no mods, apart from a bank of RCA sockets with jump leads to an adjacent blanking plate. I do however find that as the input level approaches -3dB the THD does increase, but does still remain below 0.0009%

I use a buffer comprised of a few opamps in front of the card to enable mechanical level adjustment etc.

*I note the head fi measurements are also into quite a low impedance load, I'm not sure this card is intended to drive such a low impedance.

1695584940401.png
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,373
Likes
3,318
Location
.de
A -158ish dBFS noise floor using N = 65536 FFT means -113ish dBFS worth of noise. But since that's the equivalent of an N = 32768 FFT at 48 kHz, it would be -116ish dBFS within a 24 kHz bandwidth, or -117ish dBFS within 22.05 kHz.

Considering this would be unweighted, it's probably about what you can expect out of a CS4398 (chip specs: DR 120 dB(A), THD+N -107 dB @0 dBFS; card spec: DR 116 dB(A)). Probably not quite distortion-wise, but that may be asking a bit much from a 5532.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,371
Likes
18,279
Location
Netherlands
I disagree with those Head-Fi* measurements of the Asus Xonar DX card - I have the same card and I get excellent measurements out of it.
That’s what they get as well, but once you switch to multiples of 44.1 kHz, performance is degraded. Retest at 44.1 kHz, and you’ll see the same.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,670
Likes
38,766
Location
Gold Coast, Queensland, Australia
strangely enough both of my main pcs have pci so i could fish out the cards and see if they work

but:

M-Audio Delta 410 Drivers Download​


  • Driver Version: 5.10.0.5078
  • Release Date: 2011-12-21
  • File Size: 109.22K
  • Supported OS: Windows 10 64 bit, Windows 8.1 64bit, Windows 7 64bit, Windows Vista 64bit, Windows XP 64bit
released date?

M-Audio have the drivers right up to Vista. The vista drivers can be used for 7/8 and 10. Not sure about Win11, but probably possible.

 
Top Bottom