An INA217 Mic preamp

Most of what I build is designed to reproduce sound that is already recorded, whereas this is designed to actually record sound for later reproduction. A slight departure from what I ordinarily do then, but bear with me. 🙂

It’s a microphone preamp based on the INA217 instrumentation amplifier chip from TI. The board layout is actually (another) one of my old designs that I’d managed to forget about for years but for reasons I’m not really sure about I rediscovered it and decided to rework it fairly recently. I think I did make a prototype board of the original back then but just never put it together – which is probably a good thing as I found an error in my original schematic when I did the update 😀

This design is also known as “the $5 mic preamp” (google it) since if you really pare it back to the essentials, it could be built for not much more than 5 dollars in components. My version is much more luxurious though, featuring an on-board XLR/TRS combo jack, configurable gain and phantom power as well as a DC servo and all the EMI-filtering and protection circuitry needed to avoid noise and accidents with phantom power. The only feature for this sort of amp that I have left out is the option to pad down the output with a switch – didn’t need that (and besides, no space left over anyway 🙂 )

I’ve also updated my matching PSU board which uses two small EI-core transformers to provide both the +/- voltage and the 48V phantom voltage. The transformer form factor only allows current for a single amp board, but that is OK. Originally regulating the phantom voltage was a bit of a faff, but since there is now an LM317HV with can tolerate up to a 60V-input, that was the obvious choice for the phantom supply regulator. This also means that for supplying more boards, my previous “Triple-PSU” design should be usable.

I’ve tested the preamp board with both a condenser mic and a cheap-ish Beyerdynamic dynamic mic and it seems to work quite well. I don’t have a proper recording setup at the moment though, but the sound quality is definitely good enough to warrant further experiments. I’ve made some updates to both boards and I’ll release the files when they have arrived so you can have a go yourself 🙂

Project files: STEPS clone PSU

What is it?
The board for my “STEPS-clone” single-rail linear PSU as described here. This PSU is suitable for low-power streamers, DACs, headphone amps etc. that run on a single DC-voltage rail and require less than app. 15W maximum. This isn’t really a 100% clone of the original STEPS supply (see here), but I’ve drawn quite a bit of inspiration from the STEPS so I think the credit is well-deserved anyway 🙂

Note that the transformer primary connections are hardwired on the board, so there are separate 115V and a 230V versions of the board files.

How big are the boards?
The board measures 3.95” x 4.7” (app. 100 x 119 mm)

What is the status of the boards?
The published board files are for version 1.0 which is the version I have prototyped. There are a few minor changes I could do, but it’s mostly cosmetic and it might be a while before I get to it anyway so I have decided to publish this version.

Does it use any special/expensive/hard-to-find parts?
If you can order from Mouser, then nothing here is hard-to find. If you can’t, then the only thing that might be difficult to substitute is the Murata common-mode choke and that is optional anyway 🙂

Anything else I need to know?

  • The original idea was that the board should be able to slide into a eurocard-sized enclosure (that’s also the reason for the two extra mounting holes). However, in practice this isn’t possible as the primary pins of the transformer are way too close to the enclosure walls to make this safe. My recommended enclosure is the GX1xx-types from modushop, but there are many other options. If you have more devices, you can of course use larger enclosures to hold multiple PSUs.
  • The transformer secondaries are in parallel, so with the standard Talema range from 7VAC to 22VAC, it should be possible to make the STEPS with outputs from around 3-25VDC.
  • The 2-pin header near the output can be used to connect a volt meter to display the output voltage (or it can be used for something else – your choice! :D).
  • The solder pads on the board can be used either as test points or to tap the AC or unregulated DC-voltage from the board to another PSU board for an AUX-voltage of some sort (additional circuit, trigger voltage etc.). Remember to watch the total load on the transformer and the maximum heat dissipation in all regulators.
  • You can use my spreadsheet here to calculate the adjustment resistors for various output voltages. This will show you the upper/lower limit voltages if you use a trimpot for variable output, and also the power dissipation in the adjustment resistors which you need to be careful with at higher outputs.
  • The only really tricky bit of this circuit is (potentially) managing heat dissipation if your load draws a lot of power on a continuous basis. You’ll have to balance the heat dissipation in the regulator and the pi-filter resistors, while still keeping the voltage to the regulator high enough so that it doesn’t drop out – even if the mains voltage varies a bit. A little tip can be that if your load device isn’t sensitive to output voltage, then turning up the output by app. 0.5-1V will shift some heat away from the regulator. Be sure that you stay within the specs of whatever you are connecting to the PSU at all times of course!
  • As usual for these circuits, you can use both standard and LDO (low-drop regulators). The low-drop types are normally not “better”, but can be a bit less tolerant of circuitry and load conditions so it’s actually better to stick with standard LM317 unless you have a good reason to use an LDO.
  • The only time it really makes sense to use a 3A rated regulator (LM350 or Lx1085 types) would be if your PSU is 5-7V output with a 25VA transformer. If your output voltage is higher or the transformer is smaller, the 1.5A+ current limit of a standard LM317 (or Lx1086) should be just fine.

Download design files here

Related information:
1) Read the original STEPS page linked above. Even if the circuit isn’t completely the same, there is still lots of great info about the LM317 type regulators and how to get the most of them.
2) Read the manufacturers datasheet for the regulator that you are working with. Pay specific attention to recommendations around output capacitance and bypassing of the adjust pin as there are some differences between regulator models and manufacturers here.

Note: Always read the “intro post” for additional important information about my designs.


Project files: LED tester

What is it?
The PCB files for my version of Håvard Skrodahls LED-tester as described here.

How big are the boards?
The board measures 2.0” x 1.6” (app. 51 x 41 mm.).

What is the status of the boards?
This is version 1.0 as everything (for once) worked the first time 🙂

Does it use any special/expensive/hard-to-find parts?
None, really. The 16mm pots can be bought from ebay and everything else you should be able to get from multiple different sources. If possible, I would suggest using a stereo 5k-10k pot and the fully-isolated version of the LM317. The former gives the best adjustment range and the latter helps protect against mishaps with flying test leads 😀

Anything else I need to know?

  • For information about how the circuit works, read the hackaday-post linked above.
  • Output current can be calculated as 1.25V/Rtot. For max. current Rtot = R1 and for min. current the value is Rtot = R1 +  the pot value (with the decks in parallel if you are using a stereo pot obviously)
  • There is a difference between Lin/Log pot as described in the build article, so you’ll have to decide up front which adjustment profile fits you best (or keep the pot offboard so you can change – or just build two boards 🙂 ).
  • If you want to use the “high-current” mode, populate R2 as well and short the jumper. Remember that power dissipation in both the resistor and the LM317 regulator increases with higher current. The calculations for min and max current above have to be adjusted to reflect the fact that R1 and R2 are in parallel.
  • The connection for the ammeter is required as it is in series with the LED being tested. If you don’t want the ammeter, bridge If+ and If- connections as shown in the picture. The connection for the voltmeter is optional.  Note that I have tried using a cheap LED meter from ebay for the ammeter and I had some problems with it, whereas if i connect my normal multimeters everything works fine – YMMV.

Download design files here

Related information:
Be sure to read the original post for the exact circuit description, information and tips.

Note: Always read the “intro post” for additional important information about my designs.

Linear PSUs are better…

…aren’t they? 😀

No, I don’t really want to start up that discussion here because in my opinion it’s much more complex subject than most audiophiles believe. However, one thing that is obvious is that as more and more small audio components run on single DC rails from an external PSU (streamers, DACs, headphone amps etc.), a fairly large market for aftermarket “upgrade” PSUs has opened up. Some manufacturers (e.g. Auralic) even offer separate PSUs as upgrades themselves. Well, a linear PSU is normally a relatively simple thing so why not DIY it?

Since I now have a DAC, a preamp, a streamer and quite a few other things that run on single-rail DC this seems a worthwhile project and it’s actually been on the drawing board for a while. I did have a bit of trouble getting started on the circuit and layout though, and I didn’t manage to really break the deadlock until remembered a design called STEPS by headwize/head-fi user Tangent from (many) years ago. The design isn’t up anymore, but thankfully I managed to locate it on the wayback-machine.

It’s basically a standard LM317-based PSU, but with a few tweaks added to tease as much performance as is possible out of the LM317 regulator (or one of its many derivatives). My version isn’t a straightforward copy of the STEPS, but I owe a big thanks to the the STEPS all the same. Compared to a “normal” LM317-based circuit this one includes:

  • A simple mains filter on the primary side of the transformer.
  • A snubber circuit on the secondary side of the transformer.
  • Space for high-speed/soft recovery diodes and snubber caps.
  • Space for 2+2 18mm filter capacitors in C-R-C (pi-filter) configuration before the regulator.

Everything else looks like the “high-performance” circuit variation from the data sheet of any LM317-type regulator. The onboard transformer is a 25VA Talema PCB-mounted toroid type meaning the design should be good for most applications requiring less than app. 20W power. The 15VA type transformer will fit as well and allow for mounting in a 1U enclosure, but the constraints on heat sinking and capacitor height might then be an issue.

The pictures show the completed 12V prototype for my Arcam IRdac as well as a partially completed 16V board for an Auralic Aries Mini (a recent purchase) – I’m waiting for a transformer in the mail before I can finish that and test it 🙂

LED-tester deluxe…

A few months ago I stumbled upon a presentation thread for an “LED-tester” circuit by Muffsy-creator H. Skrodahl on a Norwegian audio forum. Two things immediately occured to me:

1) I want one!
2) I think I can improve this a bit 😀

So rather than simply downloading his posted Eagle files and ordering boards from there, I started doing my own board instead. With the final result arriving earlier this week it’s time to put it to the test.

The basic idea is to use an LM317 regulator as a variable Constant Current Source (CCS) to test unidentified LEDs and confirm what currents are required for acceptable brightness – something that isn’t always easy to guess based on the published specs. I’ve kept the basic circuit intact but my modifications basically consist of:

– “Real” connectors for all connections instead of just solderpads.
– Additional outputs for LED connections to allow direct plugging in, permanent wired connections and also temporary connections via test leads/crocodiles clips.
– Space for a stereo pot to give a bit more mechanical stability.
– Optional “high-current” mode for testing constant-current LED bulbs as a supplement to just normal LEDs.
– Four real mounting holes to allow the board to be fixed to a bit of scrap metal or similar for use in a lab environment.

I need to do a bit more validation on the prototype before I publish my board files, but at least I can confirm that it works and that it is a very useful way to identify the operating parameters of e.g. LEDs in pushbutton switches.

Project files: The last of its kind…

…for a while at least 😀

What is it?
The last (and smallest) version of my EL2k buffered headphone amp using NOS Elantec 2008/2009 buffer ICs. This is the smallest version designed for 1.5″ heat sink profiles as described here. The two other versions are of course also still available (here and here):

How big are the boards?
The board measures 3.95″ x 1.5″ (app. 100 x 38 mm.) and is obviously a mono amplifier channel.

What is the status of the boards?
I’ve called this board version 1.5 as it is a redesign. Apart from the redesign work described in a previous post, the circuit is identical to the other published files.

Does it use any special/expensive/hard-to-find parts?

  • The EL2008/EL2009 buffers have been out of production for 10+ years. They can still be found and I don’t think you need to be especially concerned about fakes – there can’t be a lot of demand for these anymore – but of course no guarantees. The main risk is probably that instead of NOS parts that you get used parts that have been pulled from old equipment. This is annoying, but should be OK.
  • The heat sink profile is the same as the original, Fischer SK68, in 37mm length. Easy to get in Europe, but I’m not sure about elsewhere.

Anything else I need to know?

  • I’ve had to mount the buffers on the side of the heat sink that has an M2.5 slot and not an M3-slot. This isn’t a problem as such because there’s no need to isolate the tab, you’ll have to remember to buy M2.5 screws for mounting 😀
  • Otherwise this is a bog-standard buffered opamp circuit and there isn’t much that can go wrong 🙂

Download design files here

Related information:
Be sure to read the original posts for additional information and tips.

Note: Always read the “intro post” for additional important information about my designs.

Switched on?

Another project that’s been on hold for a very long time because I did not really need to finish it….

It’s a balanced two-in/two-out passive switch box which I intend to use partly to add an additional input to a spare amp, and partly to build a more comprehensive test setup for comparisons of sources and amps.

The signal switching is done via relays (see here for original post – oh how time flies!) and the power supply for the relays is an IRM power supply module as showed a few weeks ago. Switching of relays is done with latching button switches that I still need to wire up once the front panel is drilled and that’s about it really 😀

The relays are transistor controlled and 5V types, so while it is simple for now there is plenty of scope for adding functionality via an Arduino/ATtiny-based controller of some sort. The most obvious feature would be an IR remote control, but another thing I was originally thinking would be to add ABX-logic to try some blind testing. I sort of gave up on figuring that out, but if anyone knows how to build this in Arduino code give me a shout 🙂

Zero Audio?

Around a year ago I first tried making a music streamer from my original Raspberry Pi, a digital converter board from Audiophonics and Volumio. Apart from my unfamiliarity with Linux causing some confusion, it actually worked well and it caused me to have a bit of a mindset change. Originally I wanted to have my music stored locally on a harddrive on the playing device (a MacMini with Amarra), but since the NAS I use for redundant backup of files is just sitting there anyway, streaming was suddenly a viable option. Since then I have been happily using a RPi 3 and cheap digital converter board from ebay as a streamer to feed my Arcam DAC, switching between Volumio and Runeaudio for the software-part.

I am mostly happy with this setup, but since the RPi Zero came out a while ago I’d wanted to try using that for something similar and take advantage of the compact size. To match the Zero I bought a “TinyToslink” adapter to give the Zero an optical output to feed my DAC. It seems to work well, but it’s a bit surprising – in a good way – that something less than half the size of a credit card (excluding all the necessary adapters of course 😀 ) produces sound like this.

I have some ideas for how to case this to make it pretty, but it’s going to take a while as it’s not a priority right now. Also, the TinyTOSlink is not as sophisticated as e.g. the Hifiberry Digi+ Pro (which I have my eye on as well), and there are a few things that could be better. One of the problems is that it doesn’t do 192 kHz over optical (and my DAC will not accept that either), so I have been wondering about DIY’ing a version with transformer-isolated coax out instead – maybe later ;).

I’ll probably continue experimenting a little with the Zero and leave the RPi3 in my main system, but if you want to get a cheap streamer together the Pi – regardless of format – is a good option. And it can of course also do many other things as well (especially if you can be bothered to learn some basic Linux, which I can’t right now 😀 )

Project files: IRM Switching PSUs

What is it?
Since I first discovered the IRM-series of compact switching supplies from Mean Well I’ve grown quite fond of them. They are compact, cheap and very easy to implement so they are perfect for everywhere an “aux-voltage” is required to power non-critical circuitry. Through the different applications I’ve found for these I have managed to build up a full series of boards suitable for the IRMs.

While some of the boards can be (and are intended to be) used for “serious” stuff (to be shown later on), a very obvious application for most of these boards are as AUX-supplies for powering relays, displays, logic circuitry etc. where a bit more or a bit less ripple and noise are of no consequence, but where the compact size and low standby consumption is a real plus.

There are four board versions, suitable for the IRM modules in all versions from 3-30W output power (the 30W board is missing from the pictures as I couldn’t find the prototype when they were taken – sorry! 😀 ).

How big are the boards?

  • The 3W board measures 1.8” x 1.5” (app. 46 x 38 mm.)
  • The 5/10W board measures 1.2” x 2.65” (app. 31 x 67 mm.)
  • The 15/20W board measures 1.25” x 2.95” (app. 32 x 75 mm.)
  • The 30W board measures 1.6” x 3.6” (app. 41 x 92 mm.)

What is the status of the boards?
All of the board files are version 1.0 or higher. Some tweaks have been done after the initial protoypes for a few of them, mostly because of errors/issues with the IRM module footprints.

Does it use any special/expensive/hard-to-find parts?
No, none. Several places to get the IRM-modules them selves (Mouser, Reichelt, TME etc.) and everything else on the boards is more or less optional 😀

Anything else I need to know?

  • The modules have worse specs for ripple and noise than most linear regulators, but obviously the switching frequency is quite high (66-100 kHz depending on model), which means that passive filtering like an LC or a CRC (“pi”) filter would be an ideal way of reducing the output noise. I have a couple of examples for that which I might show later.
  • I haven’t been able to find a spec for how much capacitance the modules will tolerate on the output, but it probably should not be overdone.
  • Remember that obviously one side of the board carries mains voltage, so take the necessary precautions when working with them.

Download design files here

Related information:
Note: Always read the “intro post” for additional important information about my designs.

Class D experiments…

There are many class D technologies on the market at the moment, but one of the ones I haven’t tried (until now at least) is the International Rectifier “IRAUD7”-amps (IRF has been acquired by Infinion).

Consisting of the IRS2092 driver IC and various purpose-built FETs (many of them two FETs in a single package suitable for half-bridge designs), this is by most accounts a good-sounding and scalable class D technology. It’s also one of the few technologies where you can actually have a go at your own PCB layout if you want to. The schematics are available in IRFs published reference designs (here and here) and although making good PCB layouts for high-power switching electronics isn’t easy, it is actually possible to do.

Of course, when something is so easily available it tends to get exploited. It wasn’t long after IR introduced the designs before the market was flooded with several cheap clones, some using their own PCB layouts and some using IRs own Gerber files which are also published on the website. I had my eyes on some small (credit-card sized) boards to try for a while as they were really cheap (do an ebay-search for “IRS2092” and you’ll see 🙂 ), but eventually spotted this “luxury” version (at least based on appearance and observed parts quality) and fell in.

This build is the “low power” version with the IRFI4019 FET, but there’s also higher-power version with the IRFI4020 FET. Since the seller I bought from made a mix-up in ordering I actually ended up having a pair of each version, but I wanted to start with the low-power version. Then I might go dual-mono on the high-power boards later on if the sound quality proves it worthwhile 😀

The PSU consists of a 200VA transformer and a cheap supply PCB with 45mF capacitance per rail – mostly because that was what I had in my parts drawers. I’ve tried to keep the mechanics as simple as possible since I consider this build an experiment, but having the amp and PSU on a mounting plate simply makes everything much easier so I decided to “splurge” a little anyway :). The front panel is blank until I decide how the amp is going to be used.

Even though the pictures show the amplifiers uncabled (which they still are), I did manage some sneak listening on the modules and I am looking forward getting these into my main system for a proper test 🙂