Author Topic: Suggestion for shielded wire, inside low signal preamps?  (Read 4680 times)

0 Members and 1 Guest are viewing this topic.

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Suggestion for shielded wire, inside low signal preamps?
« on: March 09, 2016, 08:25:49 am »
Ok so first and most important I am making an order for all the passive components (IC's already obtained) on a project I have been contracted to do, I plan to make this order tomorrow. So hopefully any suggestions made are available at Digi/mouser.

I plan to do my first proto dead bug style since its all analog (I love veriboard but I dont think it will show the full potential here, just as in RF), I have been reading a lot of Douglas Self who seems to be the no BS objective authority on audio stuff and in one chapter he had talked about how prevalent cross talk issues are at low phono/mic/pick-up levels. Makes sense to me mix a low power signal just above the noise along with long wave lengths of audio...

In the final product I plan to do some shielding work and clever layout of decoupling caps between signal lines, I would like to avoid doing any shielding in the proto, cutting and soldering clad shields will slow down the process. I figure I will just run the signal lines in a single conductor shielded wire. Impedance is no issue, since transmission line wont effect me at this wavelength, so my first idea was to use that small silver coax on you find with SMA pig tails (BTW I need some of this for RF too if anyone knows what type of coax that is). Then I started thinking that is going to add capacitance to the signal chain causing LP filters, the impedance does not matter in this application but the capacitance sure does!

So I guess I need to find a cable with a single 22awg or preferable 24 and up AWG that is shielded with no dielectric, kind of like and HDMI cable with one wire and a braid that can be soldered too. Any suggestions especially that one of the big boys sells would be quite helpful!! And as mentioned above if anyone knows the name of the coax used in the SMA pigtails (used in RF, and more recently as RTL SDR adapters) that is a silver color I would also be greatfull to know what that is called!

Offline ThomasDK

  • Regular Contributor
  • *
  • Posts: 123
  • Country: dk
  • B.Eng. EE
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #1 on: March 09, 2016, 09:12:44 am »
What frequency ranges are you working with?

You mention audio, note that ordinary shielding is not effective at frequencies below 100 KHz. Mu-metal works, but is expensive. Good layout and cable management is going to be much more important than the type of cable used.

The coax used on wifi pigtails is usually RG316. 
 

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 5158
  • Country: au
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #2 on: March 09, 2016, 11:48:11 am »
What frequency ranges are you working with?

You mention audio, note that ordinary shielding is not effective at frequencies below 100 KHz. Mu-metal works, but is expensive. Good layout and cable management is going to be much more important than the type of cable used.

The coax used on wifi pigtails is usually RG316.


Mu-metal is used to shield CRTs,magnetic microphones,& record player pickup cartridges,& sometimes moving coil analog meters from fixed & low frequency magnetic fields.
It doesn't do much for Electromagnetic fields at audio frequencies.

Audio cable is quite commonly shielded with copper braid,usually with less dense shielding than coax.
« Last Edit: March 09, 2016, 11:52:48 am by vk6zgo »
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 14552
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #3 on: March 09, 2016, 04:34:04 pm »
What you're forgetting is it's not "wavelength doesn't matter here".

It's that, wavelength is weighted by the impedance ratio.

If your system impedances are on the order of 10kohms, and you are using 100 ohm transmission line (ballpark), then a 30kHz wave (\$\lambda = 10\textrm{km}\$) will see consequences for lengths on the order of 10km * (100 ohm / 10k ohm) = 100m. But not 100m but more like 1/8th of that, or 12.5m.  (10m of most cable is on the order of 300-1000pF.)

Which still isn't very small, so I kind of wonder just how high the impedance is you were expecting to use..?


The other thing you may be missing, is bandwidth:
1. Only use the bandwidth you need.
2. Bandwidth is in the eye of the beholder.

You might only be interested in 20kHz (or not many times greater than that), but that low-noise JFET preamp you might be looking at, will be hot up into 100s of MHz without any filtering or damping!

Providing excess bandwidth is inviting trouble from RFI, even with shielded cables in the system.  And here again, bandwidth can be subjective: a bipolar circuit might roll off at some MHz (as a dominant pole characteristic), but RF (in the 10s to 1000s of MHz) can be rectified as DC offset and baseband audio.  Lack of RF filtering is almost guaranteed to detect that horribly annoying dit-dit-dit sound that GSM cellphones emit.

Really, this is a lot of words just to say "add ferrite beads, caps and resistors as needed", but the fact is, they are central to the design of a quality system, and should not be waved off as second class components!

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #4 on: March 09, 2016, 05:26:02 pm »
First off yes this is audio frequency 20 to 20 at 1-4mv.

@T3, thank you for the advice, do you maybe know a ferrite mix that is good for audio applications? Impeadence in will be variable, but lets say 47KOhms in general, impedance out should be as close to 0 as possible.

Second as far as shielding, while reading a Douglas Self book I was surprised to find out in a preamp eliminating cross talk is not as easy at it seems. He list the best solution is #1 get a 4 layer PCB and sandwich each signal line in between two ground planes (well this is impossible on a clad proto), he also mentions using sheilded jumper wire or actual shields between channels. He then goes on to list good PCB techniques that keep price down, such as stacking all the power and decoupling caps in between the channels.

Point is he seems to be a no BS engineer who is quite respected, and I am making a proto ugly style ATM. So from his statements, I would think that using shielded wire in the signal line would be the fastest way to go on an ugly style proto. Now I am not going to argue wether or no sheilding is at use at audio frequencies I will instead use and example, wrap a cheap expensive extension cord driving a heavy load around speaker cable  and you will induce a 60hz hum in to it. Do the same around shielded coax... 60hz is in the audio range.

Now if I am misunderstanding something I am sorry but from his book and the practical example I used I dont understand how shielding wont help. Sooo any suggestions for signal wire?

Offline ThomasDK

  • Regular Contributor
  • *
  • Posts: 123
  • Country: dk
  • B.Eng. EE
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #5 on: March 10, 2016, 01:11:31 am »
Do the same around shielded coax...
...and you will induce a 60hz hum in to it  :scared:

It might be counterintuitive, but this doesn't make it less true - EMC is fun  :-+

A little reading:

https://www.wirelesspowerconsortium.com/technology/shielding-effectiveness.html
Shielding effectiveness of copper at 10kHz: ~0dB  :--

http://learnemc.com/practical-em-shielding
Quote
conductive materials are generally poor magnetic shields at low frequencies (e.g. below a few hundred kHz).
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1508
  • Country: ca
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #6 on: March 10, 2016, 04:26:59 am »
Shielded audio cables work perfectly at eliminating hum pickup from electrical wires all around. But the interference is electro-static, not magnetic. Therefore the shield does not block magnetic interference from a nearby power transformer. Crosstalk is also electro-static but I have never seen and I have never needed to use shielding in a compact preamp circuit.
 

Offline f5r5e5d

  • Frequent Contributor
  • **
  • Posts: 347
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #7 on: March 10, 2016, 04:53:52 am »
Whitlock at Jensen Transformers is an expert - see the site, white papers, student guide - of course his answer is a transformer more often than not but still good

Ott's book is a bible for EMC control, design of systems

http://www.audiosystemsgroup.com/publish.htm has lots of info
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 14552
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #8 on: March 10, 2016, 02:24:21 pm »
First off yes this is audio frequency 20 to 20 at 1-4mv.

@T3, thank you for the advice, do you maybe know a ferrite mix that is good for audio applications? Impeadence in will be variable, but lets say 47KOhms in general, impedance out should be as close to 0 as possible.

Yeah, at what frequency?

There is never just "impedance".  Impedance is a function of frequency!

Now try again. What impedance? ;)

Probably, around 0 ohms at DC will be fine (or no more than ~100s ohms for a line driver, or single to fractional ohms for a headphone or speaker power output).  We can let the entire audio band be "DC", that's no problem... in that case, "AC" is simply where things get interesting: a series (R || L) or ferrite bead might be used at the output, to raise the impedance at AC or RF, to dampen line resonances, ensure op-amp behavior, that sort of thing.

If nothing else, the op-amp's output impedance will rise anyway, usually being somewhat inductive (with R2R amps more exaggerated than emitter follower types).  But this can be accompanied by instability with unlucky loads (usually capacitive), so it helps to have a network to stabilize that.

Quote
Second as far as shielding, while reading a Douglas Self book I was surprised to find out in a preamp eliminating cross talk is not as easy at it seems. He list the best solution is #1 get a 4 layer PCB and sandwich each signal line in between two ground planes (well this is impossible on a clad proto), he also mentions using sheilded jumper wire or actual shields between channels. He then goes on to list good PCB techniques that keep price down, such as stacking all the power and decoupling caps in between the channels.

Well, that's rather overkill, especially when the channels are going through the air to your ears.  How much channel separation can you even discriminate? 20dB? 40dB?  It's more prominent at high frequencies (which is how electric field coupling goes), but even getting 80dB+ at that point is simply a matter of putting the amplifiers on opposite sides of the PCB (or using two boards, since they're identical anyway).

A little reading:

https://www.wirelesspowerconsortium.com/technology/shielding-effectiveness.html
Shielding effectiveness of copper at 10kHz: ~0dB  :--

http://learnemc.com/practical-em-shielding
Quote
conductive materials are generally poor magnetic shields at low frequencies (e.g. below a few hundred kHz).

This is again missing the crux of my drive here:

What's the impedance?

At 47kohms system impedance, how much magnetic field do you think will develop on any of those traces?

Approximately fuck-all!

So magnetic shielding is already way down in the noise of things you need to shield between.  The circuit impedance gets you 60dB advantage on that automatically!

The traces are still susceptible to every millivolt of induced voltage, but that's the point of high impedances: the signal voltages are that much higher, and the signal currents that much lower, so that magnetic fields in general have a small effect on the total signal.

Only the most sensitive part needs any concern.  This means placing the preamp near the connector to minimize its susceptibility.  Once the signal level is nominal, you're pretty well safe from any kind of noise (thermal or induced).

Shielded audio cables work perfectly at eliminating hum pickup from electrical wires all around. But the interference is electro-static, not magnetic. Therefore the shield does not block magnetic interference from a nearby power transformer. Crosstalk is also electro-static but I have never seen and I have never needed to use shielding in a compact preamp circuit.

Note the distinction between electric and magnetic!  This is the fields equivalent of what I'm asking: impedance.  In the near field (which is basically everything at this frequency), the impedance of free space can be tweaked up or down; that is, the magnetic or electric field can be made much more intense, and therefore |E| / |H| (which has units of ohms!) can be different from Z0.  (Only for radiating waves is |E| / |H| = Z0; or along a transmission line, |V| / |I| = Z0 when the SWR is 1.)

When the system impedance is higher than Z0 (and we're in the low frequency limit, which is the case here), a line looks capacitive.  When less, inductive.

Electric fields are largely your concern here.  Magnetic fields will still interfere, but they have to be relatively that much stronger to reach the same output level, than for a circuit at Z0.

Shielding effectiveness is still at play, so you can get induced hum on coax, even if the ground/shield experiences no ground loop (but you'll likely get that as well).

The trick with ground loop, is to "read" the voltage on a coax cable, as if it were differential.  You don't want to allow the ground/shield to float freely, but it can be connected to ground through some (R || C), so that the R opens the ground loop near DC, while the capacitors preserve a solid ground path at RF.  The input circuit consists of a differential amplifier, as it would for any other kind of input cable (like a true differential XLR jack, for instance).

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #9 on: March 10, 2016, 06:18:21 pm »
@T3

Hmmm, your last comment kind of confused me. A Phono system is unbalanced RCA's, not XLR. Im not sure what you are getting at, but are you saying I can use an instrumentation connected to the coax hot/ground, or that that is a better design choice? Im also building a mono mic pre for a friend using an instrumentation
amp becuase it is balanced XLR.

This is a bit off the topic but I have been looking at these opamps LT1115, unfourtantely there is no free samples I need two one for each line. I always thought using separate opamps was a stupid audiophool thing since there internal impedance between channels is pretty damn hi! While when reading Self's book I learned using one amp per channel is a reality that eliminates cross talk and it has nothing to do with the IC but all the passives being that close to each other in each channel. Anyways this LT1115 opamp even has a phono amp in the amp notes, it's not really how I want to design mine. I didnt choose them becuase of there "sound" when looking at the specs the noise is only about 1nv root hertz, that is much better than an NE5532/4 or any other "Audio Opamp". Since there dual channel they also give me an extra buffer to lower noise even more!

Thing is like I said no samples and there about 18 bucks for two. I will use the part if it work better than any other, but I am afraid that maybe they have a high current noise or something like that which negates the low voltage noise. Can any of you guys tell my if these are a good choice or if the datasheet is over selling the thing?

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4309
  • Country: us
  • KE7GKP
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #10 on: March 10, 2016, 07:13:06 pm »
What are you shielding "out" or "in"?
How does that affect your passive component order?
How many channels is this? 
Why is crosstalk important?  (Particularly in correlated stereo signals)
There seems to be a lot of speculative, free-floating anxiety here and very little hard facts.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 14552
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #11 on: March 10, 2016, 08:21:30 pm »
Hmmm, your last comment kind of confused me. A Phono system is unbalanced RCA's, not XLR. Im not sure what you are getting at, but are you saying I can use an instrumentation connected to the coax hot/ground, or that that is a better design choice? Im also building a mono mic pre for a friend using an instrumentation
amp becuase it is balanced XLR.

Yes, sort of.  This is better:



Note the input is grounded at RF (the capacitors might be anywhere from 1nF to 0.1uF, but probably not more), and floats resistively at low (audio) frequencies.

This is followed with a standard differential amplifier.  An instrumentation amp would be overkill, but if high precision is desired (at some expense to noise), that could be used.

Not shown: input protection, and filtering/termination.  You would probably want to filter both the IN1 and REF signals first, perhaps by bisecting R1 and R3 and adding C and TVS parts.

You don't want the input to be directly chassis grounded (e.g., metallic panel mount connectors), because that would shunt ground loop voltage to safety ground, causing the usual problem.  But you do want the RF ground handled that way, hence the capacitors.  Really, the capacitors should be connected to chassis ground, not just circuit ground.  Likely, you'd use the same for both, unless you have other reason to isolate circuit and safety ground.

The usual reason audio equipment is built with a separate circuit ground, or with "star grounding" (gack, what an abomination!), is to minimize ground voltage offsets, when multiple sources connect together.  By sensing the coax connector in a differential manner (allowing ground to float somewhat), this is neatly solved, with much better performance.



This is a bit off the topic but I have been looking at these opamps LT1115, unfourtantely there is no free samples I need two one for each line. I always thought using separate opamps was a stupid audiophool thing since there internal impedance between channels is pretty damn hi! While when reading Self's book I learned using one amp per channel is a reality that eliminates cross talk and it has nothing to do with the IC but all the passives being that close to each other in each channel.[/quote]

Beware that you'll only achieve that noise figure for fairly low source impedances.  They are using < 1k resistances in their example circuits.  There are several cautions to using this special purpose amplifier:
- Low source resistance, matched between inputs
- Input bias current is fairly low, and input current noise is quite low, but see Note 3.  What they don't tell you is that this device uses an input bias cancellation scheme, which reduces the input bias current.  (The input pair is huge, running around 2mA total (see AoE3 p.527) -- a fifth of the total supply current -- so without this, input bias current would be perhaps 2-20uA.)  That means the noise on both inputs is largely correlated, so a small mismatch in input resistance will yield an unexpectedly large error.  The remainder is the actual uncorrelated noise.
- The inputs are protected with diodes, and no internal resistors.  So you definitely can't use it as a comparator (not that you'd want to), and need to provide external protection (current limiting resistors, clamp diodes?).

Quote
Thing is like I said no samples and there about 18 bucks for two. I will use the part if it work better than any other, but I am afraid that maybe they have a high current noise or something like that which negates the low voltage noise. Can any of you guys tell my if these are a good choice or if the datasheet is over selling the thing?

Well... what's the intrinsic SNR of a phono recording, anyway?  Maybe 40dB?  You're seriously barking up the wrong tree here.  And, it's been shown that phono aficionados prefer the hiss and pop, so, are you not simply making it worse?

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #12 on: March 11, 2016, 05:13:18 pm »
Ok as far as the sheilding goes yes two stereo channels, the will be ran through two different opamps spaced aways apart, the RIAA will be passive. When reading the Self book on low signal amplification he says cross talk is a big deal and needs to be taken care of. His best answers to the problem is to stick a sheild between the two channels, which is costly in final production, or use the four layer technique. The problem is shoving ground plain between channel tracks is the magnetic currents are 3d and will just jump the 2d ground. As for prototyping I thought that using sheilded cable on clad would give good results, but if I am understanding right even soldering a piece of clad vertically between channels wont help cross talk?

This is a pretty analog product except for about 10% of it which is hopefully going to use an automatic impedance loading system using some special stuff, im not sure how well it will work of if at all. After this amp has been done I will be designing a digital version using DSP and adding USB out. I feel in the end this will be much better becuase the RIAA wont float and you can add a pop filter :).

@t3, thank you for the schematic and advice. I managed to get some samples of the LT1028 which may as well be the same product when it comes to audio, and costs the same. Im not sure what you mean by im barking up the wrong tree, I hate hiss and pop (although I was asked to build this and that is the only reason I have a turntable set up now). I had planned to use a few nE553* buffers in parallel on each channel to drop the noise floor even more, there are examples of this in Dougs book. I would think using low noise techniques would still allow pop and his since it is coming off the MM Cart and getting amplified, Im trying to use low noise to give my self head room for the passive noise.

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 14552
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #13 on: March 11, 2016, 05:36:50 pm »
Ok as far as the sheilding goes yes two stereo channels, the will be ran through two different opamps spaced aways apart, the RIAA will be passive. When reading the Self book on low signal amplification he says cross talk is a big deal and needs to be taken care of. His best answers to the problem is to stick a sheild between the two channels, which is costly in final production, or use the four layer technique.

I don't see that either is necessary.  Maybe at RF where you can get surface (evanescent) waves between sections.  That's entirely irrelevant below, like, 10GHz...

Quote
@t3, thank you for the schematic and advice. I managed to get some samples of the LT1028 which may as well be the same product when it comes to audio, and costs the same.

AoE3 calls the 1115 an "audio spec 1028".  Might be a selected/higher spec tested die.

Quote
Im not sure what you mean by im barking up the wrong tree, I hate hiss and pop (although I was asked to build this and that is the only reason I have a turntable set up now). I had planned to use a few nE553* buffers in parallel on each channel to drop the noise floor even more, there are examples of this in Dougs book. I would think using low noise techniques would still allow pop and his since it is coming off the MM Cart and getting amplified, Im trying to use low noise to give my self head room for the passive noise.

I mean, who cares about getting < 1nV/rtHz in the preamp, when the signal source is in the uV/rtHz (a hundred or thousand times worse) to begin with?  You can't recover SNR that the source doesn't even have.  I'd be surprised if a listener would notice a uA741 preamp!

The only time you'd notice a difference is when it's not playing: how much hiss there is, when the arm is lifted.  (You could solve that with a bit of subterfuge in the DSP version: reduce gain, or mute it, when no signal is present.  This would be very easy in this application, because even a blank passage, in a clean disk, exhibits measurable pop and bump and hiss.)

And as for perception, if it does add noticeable noise (i.e., the amp's noise floor is comparable to the signal's), it would probably be preferred, because phono enthusiasts prefer noise.

If you're working to a customer spec, that's fine, but I'd in turn question their choice of specs, or their priorities.

Even if you can't change anything, in order to get a better engineered system (say, including price or schedule), you should at least beware of what actually matters, from a total system perspective.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline f5r5e5d

  • Frequent Contributor
  • **
  • Posts: 347
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #14 on: March 11, 2016, 05:37:52 pm »
Quote
Anyways this LT1115 opamp even has a phono amp in the amp notes, it's not really how I want to design mine. I didnt choose them becuase of there "sound" when looking at the specs the noise is only about 1nv root hertz, that is much better than an NE5532/4 or any other "Audio Opamp". Since there dual channel they also give me an extra buffer to lower noise even more!

Thing is like I said no samples and there about 18 bucks for two. I will use the part if it work better than any other, but I am afraid that maybe they have a high current noise or something like that which negates the low voltage noise. Can any of you guys tell my if these are a good choice or if the datasheet is over selling the thing?

yes current noise needs to be considered, often rules out the 1 nV/rtHz op amps

and microphones have several different basic techs, and interface standards

raw transducers can be low Z ribbons or higher Z coils moving in magnetic fields, or a variable capacitor membrane somehow polarized which have such high Z that they nearly always have a built in FET buffer circuit

packaged microphones you buy at a Guitar Center or similar store may require a bias V to power the internal FET, or in pro mics require a balanced "phantom power" circuit


phono cartridges have a range of impedance too - low output Moving Coil types could benefit from a 1 nV/rtHz op amp, the more common, lower cost Moving Magnet phono cartridges higer Z over audio would make the current noise contribution so large that the NE5534 would be better
the low output of MC vs MM also requires an additional order of magnitude gain

Walt Jung's "Op Amp Applications" book is online for free, has a good section on RIAA preamps

I believe "passive RIAA" is a poor choice, compromising both noise and headroom
« Last Edit: March 11, 2016, 05:47:29 pm by f5r5e5d »
 

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #15 on: March 11, 2016, 07:09:14 pm »
f5, thank you I did not know jungs book was free online, along with doug self I would consider him credible and not a audiophoolish.  I had thought that using an active filter was the noisey choice? My whole reason for choosing a very low noise opamp was an idea to compensate for all the johnson noise in a passive network. I have considered tha UAF42 chip for doing active EQ but im not an active filter wiz. I mostly do digital, lately I have gotten in to analog from RF design where I use all passive filters which makes it much easier for me to implement one in this case.

@T3 im just trying to do something like the 02 headamp, which is to design something at the $150 or hopefully less price mark that objectively measures well! To me the graphs when hooked up to a dScope is what matters, not what aficionados care about. Im working with an audio engineer that works for a big company on every receiver that starts with a D and getting help and tips from someone who designs pro gear and acually sells alot of it to professional musicians balh blah blah. The idea of this concept is an entry level PreAmp that wont break the bank which measure as close to perfect as possible and is easy to use, hence the use of a digital loading stage of the cart, I dont want people screwing with jumpers for capacitive loading etc... This is mostly geared twords kids, hipsters getting in to the vinyl resurgence.

Lets face it most 3000 dollar preamps are shit with foolish component selection, there caps probably cost more than an LT1115. I have a selection of entry level pre amps from 15-100 bucks and some are designed better than others but none of them past objective measurement tests very well. For me this is about designing a product in the right place at the right time with the right contacts, to lead to better things. I like the viynl experience but it sounds like junk, hence wanting to do a DSP amp as the next product. The only ones Ive seen that use ADC/DAC combos with DSP are 1K plus which is ridiculous a two PSOC5 chips at 18 bucks could handle the RIAA plus pop/click smoothing! and then what maybe 40 for the ADC/DAC and an FX2/FX3 for USB. This is a terrible market and I hate it honestly.

What is funny is if you guys are correct about the shielding, which doing RF I would tend to agree with you.. that means I cant even trust these "No BS" audio designers like self and jung. When I read some of there stuff I think wow I had no idea that mattered a low frequency but these guys are the industry leading engineers.... As to argument about sheilding I only tended to by in to it becuase self says it is more about the low signal levels and this stuff is not an issue in line level or power amps.

Offline f5r5e5d

  • Frequent Contributor
  • **
  • Posts: 347
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #16 on: March 11, 2016, 07:29:29 pm »
http://www.ispra.net/audio/preamplifier/thorsten-loesch/thorsten.html imo Thorsten is actually a solid engineer - but with tweakazoid, Audio Guru bombast leanings

today I would look at TI's newer fet input device OPA827, or the cheaper opa164x

the discrete fet is 2nd sourced by Linear System, a budget option could be one or more BF862

the back-to-back polar electrolytic DC blocking is dated too - Batemann's "Capacitor Sound" series showed that bi-polar Al electrolytics with full thickness oxide on both foils gives lower distortion than the back-to-back connection
« Last Edit: March 11, 2016, 07:32:47 pm by f5r5e5d »
 

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4309
  • Country: us
  • KE7GKP
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #17 on: March 11, 2016, 08:29:09 pm »
That kind of extreme shielding between audio-frequency channels (even UN-correlated, NON-stereo channels) is unheard of in the professional world. It is the lore and folly of audiophools, IMHO.  it is trivial with ordinary circuit layout methods to achieve MUCH better channel separation than the stereo signals you are going to be processing. Particular if you are talking about audio from a 45/45 (Blumlein/EMI) stereo phonograph disc. 

IMHO, passive RIAA equalization is an audiophool artifact of very early days of semiconductors and has no advantage over more robust modern circuits.

NWAVguy (http://nwavguy.blogspot.com/) has probably the most cost-effective, true audiophile (pleasing even audiophools) headphone amp design out there.
You would do well to study and emulate it, or at least retrieve it before the website disappears as the author seems to have.
http://spectrum.ieee.org/geek-life/profiles/nwavguy-the-audio-genius-who-vanished

FYI: https://www.gearslutz.com/board/geekslutz-forum/ is a more specific audio circuit design forum where you can interact with some pro audio industry professionals.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 14552
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #18 on: March 11, 2016, 09:30:00 pm »
Just because someone is a professional something, doesn't mean they are right about every little thing.

The biggest problem is measurement.  There is a tremendous about of misinformation out there about EMC (electromagnetic compatibility), even up to manufacturers' app notes (which are generally rather poor on their own, not just in regard to this aspect).  There are many documents which tell you patently false things (like adding ferrite beads in series with shield ground connections), and others which indirectly illustrate the false premises they set out to prove (like "minimize switching loop inductance").

None of these false documents would exist if their authors took the time to analyze the variables in their circuits, and measure and quantify results.  But measurement is hard -- not really so much that it's hard to do, because you can rattle off a hundred things and have them measured and documented in a couple hours, if it's something simple to change (like changing lumped component values, as opposed to, testing a hundred PCB layouts..).

So what it really comes down to is:
- A document hides the character flaws of the author. You see what the author/editor wants you to see (assuming one or the other even cares enough to achieve this..).
- If you met the person face-to-face and saw and discussed their work, it would be immediately obvious if they were methodical or not.  At least, in a relative sense.
- Above all, engineers are lazy.  It isn't surprising, but it is unfortunate, that most do not take the time to actually test their knowledge.

And applied to the present project, that's exactly what I'm saying, and Rich and f5 and such.  There's no point in super-low distortion, or noise, or isolation, on a project like this.  Who cares if it measures better than the competition, if the competition is still e.g. 5x better than the source material?

And yeah, if you think there's a market there for "measurably better", at "competitive pricing", and you'll make back the NRE and sell enough quantity to get the unit cost down there, then go for it.  But you'll get justifiably better performance (meaning this time, in the marketplace) with a cheap jellybean op-amp and a slightly nicer circuit, saving those $2/unit or whatever.

Or still other ways to solve the exact same problem: just out-and-out lie about it on the marketing literature.  Make it 0.1dB quieter than arbitrary competitors A, B and C.  It's still good enough for the job!

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline rwgast_lowlevellogicdesin

  • Frequent Contributor
  • **
  • Posts: 565
  • Country: us
    • LowLevel-LogicDesign
Re: Suggestion for shielded wire, inside low signal preamps?
« Reply #19 on: March 11, 2016, 11:00:43 pm »
So I have read NWAVGuy quite a bit, my first all analog project about 5 years ago was a head amp based on some tangent stuff with NWAVGuys ideas to bandaid up the amp.

Like him Im looking for good measurements and I have a decent technics turntable with test records but thats rubbish, I have developed a home test system to InverseRIAA FLAC files then feed them in to a phono pre and then in to a motu 196/24bit audio interface. I analyze that Using smaart and some other pricey software for THD etc. I also use a 12bit USB siggen/scope, and a 25mhz 8 bit DSO with FFT. In the end I plan to publish test using a viynl system called adjust plus and dscope measurements.

While an LP has a loud noise source, a FLAC does not. Ive been doing this project using "Test driven development" where you come up with clear testing methods and parameters before you even put any circuits or sims of you idea to test. I guess im hoping if I can make this thing play a FLAC perfectly it is going to have some good numbers, especially better than numbers straight from a phono cart. This to me is the amplifiers capability, if there is anything sounding horrible well then we know its not the pre amp. Records sound like shit period. I also believe that 5532's are plenty enough for a good phono pre anything better is basically ramping the measurements and overkill. BUT maybe this overkill will make a difference with a brand new LP on a good table with a good cart that has proper alignment.

I dont know, the LT's were free if I gain nothing ill just pop em out and use 5532 or the JRD4552 used in the 02 amp. In my opinion the phono input on my stereo sounds just fine, but that is subjective. I also almost exclusively listen to podcasts anyway so what do my ears know ;)


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf