MINIRIG

assembly using inexpensive YuanJing brand TDA8920 board, with some minor tweaks.

ZOG — Tue, 13 Jul 2010 11:19:10 +0000

I recently got a YuanJing brand low cost TDA8920 based board from ebay. This is a fairly basic low frills build but worth documenting for those interested.

I’d already wired up the case for my second ConnexElectronics TDA8920 board when I managed to fry the board. Out of curiousity I decided to check out this lower cost board that was going cheap cheap ($19USD + $9USD postage), as I could use the transformer I’d already mounted in the case as they are similar units.

This is the board on ebay:
TDA8920 Class D Amplifier board/ audio Power AMP 2X100W - eBay Power Amplifiers, Amplifiers, Home Audio, Electronics. (end time 19-Jul-10 02:14:45 AEST)

It has a smaller heatsink than the connex module, and smaller PSU 4700uF caps.

Initially it sounded OK, but not as good as the Connex module, quite a bit more grainy.. also interestingly it uses 5W power even when silent, and hence (with the smaller heatsink) runs a lot hotter.

Tracing out the circuit it appears it is pretty much the reference implementation of the TDA8920 circuit, with a volume control tacked on. Though the PCB mentions 0.680uF caps, they have used values 0.470 as per the reference circuit

Some minor tweaks I have made to the unit include:

remove the input capacitors, and put the inputs where the capacitors were. This cuts out the 0.47 coupling caps, the volume control, and some 5.6K resistors. Also means I have to make sure anything I plug into it doesn’t have any DC offset! This has really cleared up the amps sound quality a lot.
took the 10000uF caps from the connex amp and wired them in parallel with the existing 4700uF caps. Seems to have more solid bass… but hard to say.
took the 11uH inductors from the Connex amp and replaced the 33uH inductors (which are in metal cans). After comparing the circuits and discussing with Col also replaced the 0.470 caps output filter caps with 2 in parallel (0.95uF) to match the values as per the Connex amp. Hard to tell if this helped either.

The overall result now sounds better. Not incredible but quite OK for a modded cheapy. Interestingly this unit doesn’t seem to have what I call the “Diana Krall” bug (faint fizz audible on very soft piano notes) thats been noted on other TDA8920 based modules.

Other parts used for the build (besides the parts stripped from the Connex TDA8920 module) include:

160VA 18+18 toroidal transformer from Altronics for $52.50.
ABS plastic instrument case from Altronics for $14.95.
gold RCA sockets from Altronics for $3.45 each.
heavy duty banana sockets from Rockby

The the ESP groundloop protection has been added to the circuit to ensure quiet operation.

Minirig Symetrix 420 ConnexElectronic TA2022

colren — Mon, 10 May 2010 23:55:55 +0000

1.6kg, 100W per channel, Class-T with SMPS.

Just finished this amp that some may find interesting. I wanted to build a 19″ 1RU (rack unit) Class-D power amp that would be light and compact. Using the ConnexElectronic TA2022 amp module, the SMPS250 PSU and the speaker protection module. It was a bit fiddly as I had to cut 5mm off the heatsinks and replace all the big caps with 35mm ones. (click pic)

tweaking the Dick Smith Electronics DSE A2667 bookshelf speaker

ZOG — Mon, 03 May 2010 13:47:42 +0000

I recently got some Dick Smith Electronics A2667 bookshelf speakers cheapish ($20AUD) off Ebay. For a bit of fun I thought I’d see how much better I could make these low cost “made in china” (though not quite white van) speakers sound.

Dick Smith Electronics is an Australian retail electronics chain who used to have a more hobbyist/kit orientation, but now more known for selling laptops, TVs, home theatre rigs etc. They have a bunch of lower cost “home brand” gear, some of it quite, well, cheap, but occasionally of quite reasonable quality.

The A2667 speakers were sold 6 or 7 years ago and I had read on various forums that for $70 bookshelf speakers they competed well with speakers under $300, and could be made even better by upgrading the crossover.

They are small 2 way bass reflex speakers with a 5″ woofer tweeter is 1″ fabric dome. Rated at 100W with 86dB efficiency.

Unmodified they sounded quite good (for the size and price) but a bit muted in the high end, and a slight graininess to the sound. But for their size and price they have quite a big and pleasant sound.

Running up ARTA and doing a frequency response I saw a huge notch in the frequency response at 8Khz, presumably the crossover point. The crossover turned out to be a simple first order crossover with a 2.2uF elecrolytic capacitor with a 0.6mH iron core inductor, as well as an lpad for the tweeter.

I was measuring this in a small room with a lot of stuff around so the low end measurements would be all over the place from reflections..

Hoping the notch was from lazy crossover design rather than driver limitations I plugged in a few numbers into some online crossover calculators and decided replacing the 2.2uF electrolytic with 4.4uF (two 2.2uF MKT capacitors in parallel) would close the hole, with parts I had on hand.

Measuring again and the 8khz notch had been greatly reduced confirmed by listening and hearing a lot more detail in the high end.

Next i replaced the inductors with 0.56mH aircore inductors of a higher gauge of wire. I hadn’t measured any THD and this didnt make a measurable effect on the frequency response, but the sound really opened up with more clarity and detail, and the previous slight graininess was gone.

I could probably get even better results with some simulation of the cabinet and drivers etc, though it might get into diminishing returns territory.

The net result of this is these cheap speakers now sound great, detailed and pleasant to listen to with a lot of high end detail. I’m impressed how much some relatively simple component upgrades helped on these speakers.

old capacitor and new capacitors:

old inductor and new inductor:

rear of the speaker:

ConnexElectronics TA3020 V3c amplifier module.. built into a ugly but functional steel case.

ZOG — Thu, 08 Apr 2010 09:26:11 +0000

I recently ordered one of these modules after being very impressed with the sound of an earlier version of this module I borrowed from Col. It’s an amp module using the Tripath TA3020 IC, which can do up to 110W at 8Ω, with max. 0.1% THD, though I’m more interested in the 50W per channel it can do with just 0.02% THD.

This version of the module has a built in bridge rectifier and power capacitors, it just needs AC power.

I ordered it with the dual 18,000 uF Elna caps option, no fan.

The choice of MOS-FETs was added to the page after I ordered, so I guess I got the default.

Being eager to just Get The Thing Up And Running Right Now I have quickly cased it up using an discarded steel power supply case I found in a recycling bin a while back. Luckily this steel case already had all the grounding points, and a medical grade EMI filter as well, so I already have the starting point for an OK build. It’s actually nearly the perfect size for this project, and I repurposed some of the existing holes in the chassis so I didnt have to cut new ones.

As well as the amp module itself I got a soft start module from the same vendor as it is recommended as I’m using a 500VA toroidal transformer, to avoid zapping things when I turn the unit on.

I’ve chosen the 35v-0-35v 500VA transformer as I’ve been told 35v is about the maximum I can drive a TA3020 before it needs a fan to cool it, and I just want to use the heatsink it ships with.

I’ve implemented the ground loop protection as per the Elliot Sound Products article (thats the diode bridge in the middle).

Just using 600V 14AMP cable to connect the speaker outputs, which are poking through unused fuse holes on the box affixed to a bit of masonite for to insulate them from the chassis.

Luckily a phenolic RCA dual socket I had lying around from a long abandoned project fitted one of the DC holes with the screw holes even matching up, as I dont have the tools to cut the steel case.

outside of the case (before I cleaned it and stuck the rubber feet on):

the sound
Ok the case is a bit rough around the edges, but I confident it’s wired properly and safe enough to sit down and have a listen - I tried it first with a CD walkman and some crap Tandy minimus speakers and it worked first go and no smoke escaped… so out to my “cheapowave” Econowave speakers, being driven by an AlienDAC DAC into an Elliot Sound Products p88b preamp.

It sounds GREAT.. really firm, powerful bass, while keeping high end space and detail. A little bit more volume and possibly bass definition than the older TA3020 module (but I am feeding it more voltage via a much larger VA rated transformer).. and heaps more bass and detail than my ART SLA-1 100W RMS Class AB power amp.

Almost dead silent, some white noisy hiss audible if I put my ears right up to the horn of the econowaves, but nothing at normal listening distance. Note I suspect the ART SLA-1 is a bit quieter.

A bit later I spruced up the build a bit, after initial testing showed it to be OK:

added gold plated rca sockets
illuminated on/off switch (12v switch controlling the soft start module)
bolted the ground loop breaker and toroidal transformer to the chassis after drilling holes for them.
snipped the simple header wires provided and soldered in some shielded audio cable instead.
drilled some holes and mounted the two PCBs with m3 nuts and bolts and nylon spacers. The first attempt caused some ground loop hum, but replacing the metal nuts and bolts with polycarbonate ones from Altronics fixed that issue. Cristi from Connex says “the amplifier mounting holes are isolated from GND except the middle one between the inductors, and that can be used as the GND star point“.
Some rubber feet.

still a pretty spartan hacky functional build.. but soundwise definately up to scratch.

Note: this an edited version of some posts to AudioKarma forums..

Orbital Live at Billboard

colren — Sun, 21 Feb 2010 20:00:02 +0000

An amazing concert. Very lucky to be able to see Orbital live in such a small venue.Iv’e put up a few photos taken by Bunny from the show and added a few extra of the Orbital live rig (click pic)

Minirig K9 - ESP P88 preamp build

colren — Thu, 28 Jan 2010 03:50:32 +0000

This one has been on the back burner for a while as Iv’e assembled it a few times and then discovered ground loops. Now all cured (click-pic)

ART 4 Way 24db Crossover ESP P125 Retro-fit

colren — Wed, 27 Jan 2010 03:47:50 +0000

Recently, I have been busy working on a new minirig PA which was going to be called v7 and was almost compleated as a 3 way active. Now Iv’e decided to skip v7 altogether and move straight to V8.

Minirig V8 will be a 4 way active. Over the Australia Day weekend I finished retro-fitting this pair of 4 way mono crossovers which will be at the heart of the system. (click pic)

cheapowave speaker build, MK II - econowave

ZOG — Tue, 26 Jan 2010 13:25:50 +0000

After experimenting with various combos of crossover values, although the MK I cheapowave using the Pyle PDS221 compression driver sounded OK.. the high end was never really impressing me.
Note: this post continues on from cheapowave MK I postwhich documents building the cabinets etc

After reviewing the Econowave thread on Audiokarma yet again I ordered in a Selenium D220Ti-8 from Parts Express in the USA. The shipping cost almost as much as the drivers themselves.. but although I could probably find a more than acceptable replacement from the P.Audio drivers from Cannon Online, I decided to go with a combination that already has shown so many good results for others, plus the hard yards on the crossover circuit has already been done by people with way more expertise than myself.

The Selenium drivers arrived, and the immediate impression is that they are bigger then the Pyles.

After screwing them in and measuring them… they are also much louder than the pyles.. on paper 7dB louder but I think a bit more. This is why the high frequency compensation techniques don’t work as well on the Pyle PDS221.. it just makes the driver too quiet to be usable with the Pioneer woofer…

After a bit of measurement and messing about with speaker workshop (and a few mis-tries, as I don’t have anything calibrated so I have to measure relative to the current setup, and sometimes something changes in the measurement rig by accident and knocks everything out of whack) I went for the standard low efficiency woofer econowave crossover, but with some attenuation on the compression driver via a 22 ohm resistor in parallel, and sticking with the first order low pass crossover that I’ve found seems to work OK with the Pioneer woofer.

update: this diagram not right, it leaves out a 22 ohm resistor in parallel with the selenium comp. driver

After being resoldered lots of times in different configs, the bits of cardboard used as the prototype crossover starting to get a bit worse for wear.. also I’m combing a few components to make up values as well, so the crossovers look a bit rough until I re-make them (but they work just fine):

SOUND:

The main thing is they now sound awesome.. before the best I could get was “not bad”.. but when playing “ Emmeleia” by Dead Can Dance on the new layout and hearing little echoey bits in the corners I hadn’t hear before I knew these things were ready.

Since then I’ve just been listening to them… another good sign, but I just did a round of measurements to post up here.

white: calculated output
black: measured output
pale green: Selenium output, no crossover (33 uF protection cap)
yellow: Pioneer woofer output, no crossover
(click to get link to bigger version)

Interestingly while I was running up the calculated output I found a small improvement I can make.. currently I have the compression driver hooked up with negative to positive as in the “standard” Econowave layout, but putting back the comp driver to have positive hooked to positive yields a slightly smoother curve at the crossover point in Speaker Workshop. Given how closely the simulated response matches the measured response I’ll change the polarity of the horns.

Just as a reference I hooked up the recently re-capped Alesis Monitor ones and measured them with the same mixer settings and measurement rig, here are the response curves superimposed - each time the mic is about 1M from the speaker and midway between the HF and LF driver.

Black: the econowave
White: the Alesis

Except for rebuilding the crossovers more neatly onto a bit of masonite or similar rather than the messy bits of cardboard this stage of the speakers are complete.. and they sound great.

Ideas for MK III?

Although the Pioneer A30GC50-52F-Q-1 sounds fine the response has quite a lot of dips and bumps in it, especially at or around the crossover point.. which has made getting the crossover right a bit fiddly. Although first I need to measure it without the effects of the room on the curve. Then perhaps I can see which 12″ woofers are available which have a smoother response curve up to about 2K which will work well in a 120L bass reflex box… but right now they sound just fine so no urgency at all….

Addendum - rebuilding the crossovers and putting them in the boxes
Now they sound just fine I rebuilt the crossovers onto some bits of MDF, and mounted them using cable ties to the bracing inside the speaker. To stop them rattling around I’ve damped the join with a big pile of hot glue, as well as all the components being stuck to the board with hot glue as well.

minor tweak : re-capping the Alesis Monitor One (MK I) studio monitor speakers.

ZOG — Tue, 26 Jan 2010 03:58:16 +0000

These speakers have been a favourite of mine for ages. I got them over a decade ago second hand for $600. Given they are now at least 12 years old and probably more I decided to replace the capacitors in them. As stock they are KSC 50V non polar electrolytics.

First I removed the crossover and checked the capacitor values:

Now the values are 5uF, 16uF and 24uF.. these aren’t commonly available values so I made up close matches of these values by combining various capacitor values from the polyester and polycarbonate capacitors available from Rockby Electronics.

C1: 16uF - made up using 2 x 6.8 uF and 1 x 2.2 uF
C2: 24uF - made up using 2 x 12uF
C3; 5uF - made up using 4.4uF and 0.22 uF.. just to be obsessive as one measured a bit low I added an extra 0.1 uF so they matched more closely.

Here is the crossover circuit as mapped out by tracing the PCB.

	Left Speaker		Right Speaker
	original value	replaced with	original value	replaced with
C1 (16uF)	18.7	16.1	18.3	16.4
C2 (24uF)	25.7	24.1	26.8	24.7
C3 (5uF)	5.4	4.8	5.3	4.8

C2 especially has drifted up in value to be out of spec on each side, 17% and 14% off respectively.

The crossover looked a lot bulkier after the replacement:

Measurement: using a quick and dirty measurement setup I took some before and after ARTA impulse measirements, here they are plotted using Speaker Workshop. There does appear to be a measureable difference after replacement (right speaker measured).

(click to get higher res image)

Listening: I haven’t really had much time to listen to them in detail, on quick eval they feel less shrill but that could just be that I want them to sound less shrill. But at least with these capacitors the drift in value over time has been arrested.

I used a dab of hot glue to stick the capacitors in place so they dont vibrate, as they are now assemblages of multiple capacitors protuding from the PCB somewhat.

Also hot glued the corners of the terminal plate for a bit of strength, the plastic corners had become crumbly with age and had just fell off when I unscrewed them, though not looking as neat the hot glue will be an adequate replacement.

minor tweak : re-capping the JPW Sonata bookshelf speakers

ZOG — Mon, 11 Jan 2010 11:05:19 +0000

I while back I got some JPW Sonata speakers from EBAY for $67 AUD. They are nice “entry level” English bookshelf speakers of reasonable quality…

Laurie Anderson is sitting on my bed singing to me this evening. Earlier Tom Waits popped in.

Basically, I am gobsmacked by these little sealed boxes. I don’t know if it’s because they are in a small, quiet bedroom, mounted on a solid wall, or because they have an Audio Innovations amplifier powering them. They are sublime and emit this beautiful warmth and richness that makes me not want to leave the room.

I’m not quite as impressed as the above reviewer, but they are very nice little speakers for the size and price, especially for acoustic music.

And the sealed enclosure means they mate very well with a subwoofer.

Opening the back the crossover is a simple first order consisting of a 3.3 uF electrolytic capacitor and an inductor. Presumably the resistor is to attenuate output to the tweeter.

As the speakers are possibly 18 years old the electrolytic could be getting old, or prone to drifting in value, so a simple upgrade is to replace the capacitor with a better type that wont drift over time.

Interestingly removing the capacitors and measuring them they were 3.5uF and 3.6uF respectively.. still within the rated 10% tolerance. The caps are 50V ELCAPS.. googling reveals their current price of 40p in the UK (72c ?).

An inexpensive but quite reasonable upgrade was some Vishay MKT 1822 capacitors from Rockby

Can I hear the difference? Hard to say. I think they sound better… but then I just upgraded them so my brain “knows” they should sound better.. But they wont be susceptible to aging like electrolytics.