NJM4580V MSOP in pre for recording?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Does it matter?

No it only matters to the extent that I'm curious. My own curiosity matters to me (of course) but what matters to people is entirely a subjective thing.

If you must know, changing the OA, and caps, significantly changes spaciness, 3D and room in the recordings.

Who says so? And to what opamp as a replacement?

About the SMPSU - it will be down to grounding and shielding/filtering. Which in a consumer product is normally strongly curtailed by lack of space. If its a portable recorder it will probably not be possible to maintain portability (i.e. compact shape) and fix this. What's the model incidentally?

Based on my experience I'd say fixing up the SMPSU noise would improve those 'audiophile' aspects better than swapping out that opamp.
 
yeah, power supply noise is about steel boxes or walls, ferrite beads or toroids on wires, disk capacitors in the right places to kill RF, etc. Look at the Dynakit PAS2 preamp- two steel boxes, one for the power supply, an inner steel box for the sensitive bits. Look at an Ishida food weighing scale. They are measuring microvolt signals with a cycle time of 1 package a second some times. The switch mode power supplies are around the corner of a big steel box totally away from the measuring circuits. There are toroids on the DC wires, and also toroids wound into the AC input lines.
I made a big improvement in my disco mixer just by getting the power supply out of the steel op amp box entirely, and buying a wall transformer. I also put toroid chokes on + & - DC in, and put disk caps across the wound caps and at the op amp socket, and across the feedback resistor. Changed from 4558 op amps to 33078 or NJM4560 (both sound about same). Way quieter. There is no plug in solution to noise control. My disco mixer had a competent circuit designer, but the packaging guy was too sensitive to marketing and cost control, that part of the design was stupid. Sensibly designed packages that don't have glitz don't make it to the store shelf these days. Those kinds of products are limited to boutique stores with 1000% markup in megacities - or the internet. Look at repackaging to get rid of your SMPS noise, however hard it is. NJM4580's are a fairly competent op amp. Op amp noise is a white hiss, not any particular frequency, unless you suck too much current out of one like the LM4560, that will current limit snappily.
 
Last edited:
Administrator
Joined 2007
Paid Member
What are you answering? "Package / Case:SOIC" !?

Well I wasn't sure what you were asking :)

So my link gave availability of SOIC package from Farnell and also full data sheet for that IC from an embedded link within that Farnell page.

If you are after something that would sound "better" I would suggest you at least consider and try the OPA2604 FET opamp.
 
No it only matters to the extent that I'm curious. My own curiosity matters to me (of course) but what matters to people is entirely a subjectiv...
Based on my experience I'd say fixing up the SMPSU noise would improve those 'audiophile' aspects better than swapping out that opamp.

Tascam HD-P2

And yes, SMPSU needs to be taken care of too! All constructive ideas are more then welcome!

PDF of SMPSU and pics of board
Dropbox - Tascam HD-P2.zip - Simplify your life
 
yeah, power supply noise is about steel boxes or walls, ferrite beads or toroids on wires, disk capacitors in the right places to kill RF, etc. Look at the Dynakit PAS2 preamp- two steel boxes, one for the po.....

You surely have something here!!

Take a look at the uploaded PDF and pic, and give your best go at a solution.

Here is a recording sample, without mics! So just the noise! Take it into any audioprogram, and see in analyzer which name I forgot, but a graphic image of the sound, and black background. I'm posting it here too.

Dropbox - hissnoise.zip - Simplify your life

hissnoise.jpg
 
Last edited:
Well, I can't read the units on your trace, but if the period of that nasty spike is .1 millisec to .01 milliseconds, I'd say that is your switcher. One thing you can do figure out what voltages you need (+- 12 or so for op amps, what else, power amp?) is disconnect the base of the switcher transistor, to shut off the switching, and bring in your dc from linear supplies outside. I'm using a wall transformer going into my steel disco mixer, with toroid chokes on entry (25 turns) disk caps across the supplies (.1) in addition to the 1000 uf supply electrolytic caps, and two zener diodes as regulators since all my load is op amps. You have to use resistors between the incoming and the zeners to keep the zeners from burning up. I'm using 1n5344's with 22 ohm resistors, for +-7.8 VDC. the wall transformer is an 18VDC one from a race car set, bought at Salvation Army for $1. If your device is not in a steel box (not stainless steel), you may have to encapsulate it. Signal ground (ring of the RCA jacks or back shaft of phone plugs) is the point between the two zeners, where the line touches the point of the other zener.
 

Attachments

  • schematic-wall-supply-regulator.pdf
    82.8 KB · Views: 49
Last edited:
All constructive ideas are more then welcome!

In this case, something constructive escapes me. I'd abandon it and buy something a little better engineered. Seriously.

I examined the layout of the LT switchers. They run at 1.25MHz according to Linear Technology's datasheet. Not an SMPSU for the faint-hearted at all, given that switching frequency. Tascam's engineers saw fit not to follow LT's recommendations on layout (making the smallest possible loop size).

Your wav file shows fairly broad-band noise, but the single largest frequency component is at 46.3kHz. Any idea what it might be?
 
In this case, something constructive escapes me. I'd abandon it and buy something a little better engineered. Seriously.

I examined the layout of the LT switchers. They run at 1.25MHz according to Linear Technology's datasheet. Not an SMPSU for the faint-hearted at all, given that switching frequency. Tascam's engineers saw fit not to follow LT's recommendations on layout (making the smallest possible loop size).

Your wav file shows fairly broad-band noise, but the single largest frequency component is at 46.3kHz. Any idea what it might be?

I would love a Sound Devices 722 - but it is too heavily priced!
And I know, that the Tascam can be just as good, with a few hacks - I just need you guys to figure out which! It is done on completely new decks, in the us for some money..
Here is a few leads: "The Basic MOD FR2LE sounds better due to replacing the bi-polar chips in the preamps with Field Effect Transistor (FET) chips. All the chips in the signal path up to the A/D chip are upgraded with the best chips my ears like and test gear shows work well. Some resistors are also upgraded.

The stock chips have Total Harmonic Distortion (THD) down about -91dB while the ones I use have THD down about -107dB . They also have 30% less noise than stock chips. FETs sound more like a valve than a switch and many users prefer this difference. For most listeners it is most audible with soft signals or signals with a really wide dynamic range. If you own a Hi Rez playback system it is clearly audible at any signal level. Listeners who value imaging and the ability of a recorder to capture a 3D soundfield quickly recognize the sound of 120dB SFDR FETs in the signal path.

The Super MOD FR2LE sounds better not only due to replacing the bi-polar chips with FET chips but also due to an improvement in the capacitors used in the signal path. The low grade electrolytic caps the stock machine uses have artifacts down at -85dB while the ones I use push that down to film cap levels of -125dB. With an input signal down at -40 to -55 dB you need input stage artifacts to be as low as possible as the preamp amplifies both the desirable signal and the undesirable artifacts. All capacitors in the signal path to the A/D chip are upgraded. The Super MOD does bypass the ¼" inputs and its mechanical switch. This allows me to use a higher grade capacitor in the circuit and that is why I do it. The Super MOD offers a significant improvement in detail and better nuance and depth of texture for acoustic instruments, FX, nature sounds and voice. The maximum gain of the FR2LE Super MOD is about 54dBV.

All op amps are changed in the input path along with the input capacitors. Resistors are changed to give a better gain stage for microphone inputs. The purpose of this modification is to open up the sound for a better soundstage. The transparency modification gives excellent transient response which translates to better detail throughout the audio frequencies. Bass has tighter definition the midrange is more natural and open and the high end is clean and detailed without harshness. the noise floor is lowered due to the resistor change and the low noise of the new op amps.

The one that started it all for Busman Audio. Improvements are huge with this upgrade, bringing better detail and soundstage to this interface. This modification has been compared to some of the better portable pre amps on the market. This modification changes the op amps and capacitors in the input path including the RCA inputs. The capacitors for the power supply and phantom power are also upgraded to give the user the best quality this interface can offer."


Is there a simple way to alter that loop size, or gain something good?
I have no idea about what the broadband noise is all about!
 
Last edited:
mixing switchers and audio requires real engineering. Lacking all that I'm forcing my stuff to linear Power supplies while you can still get surplus transformers.
I don't know what you are hearing the difference between 90 and 100 db down on harmonic distortion on. My speakers are similar in drivers to SP2's that are specified 30 to 40 db down on HD. They are the best I have ever owned. I don't recall seeing any other brand of speaker that plotted HD against frequency for 2nd and 3rd harmonic at 1 Watt.
I do own one amp with a switcher power supply, the CS800s and it sounds exactly the same as the Djoffe bias mod TIP mod dynakit ST120 amp with linear supply. The CS800s has all the tricks I mentioned. Separate steel box for the power supply, chokes or beads in and out, disc caps. If you are capable of reducing switch noise from a power supply to 87 db down, you can go in competition with audiopower.it that is selling a switcher package for amps that they quote that noise figure for. They charge E160, a lot more than the surplus transformers I am buying.
 
Last edited:
mixing switchers and audio requires real engineering. Lacking all that I'm forcing my stuff to linear Power supplies while you can still ge...chokes or beads in and out, disc caps. If you are capable of reducing switch noise from a power supply to 87 db down, you can go in competition with audiopower.it that is selling a switcher package for amps that they quote that noise figure for. They charge E160, a lot more than the surplus transformers I am buying.

Could you give it a go, and sketch where to put what on the diagram I posted?
 
Here is a few leads: "The Basic MOD FR2LE sounds better due to replacing the bi-polar chips in the preamps with Field Effect Transistor (FET) chips.

Given all that out-of-band noise sloshing about in the design, I definitely endorse this. JFET inputs generally sound better than bipolar in such hostile environments. The THD measurements in the datasheets though have precious little to do with this.

Is there a simple way to alter that loop size, or gain something good?

Nope - fitting linear power supplies is your best solution, and that's not a simple mod.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.