Alesis RA-100 Amp Measurements and Mods

[SonsOfThunder]

Okay...

There seems to be some interest in this so I'll start here. Who can host the images for me? I should have some pics and some pdf graphs soon. I know I have graphs at work on the unit in its stock state but I don't think I have graphed the sweeps since I modded (two iterations).

These things are actually a decent basis for an acceptable amp, after mods. If nothing else, you can use the chassis, heatsinks and power trafo, and etch and stuff your own design here. You'll spend more than the cost of a used one of these for those parts alone.

The design is all discrete (no ICs at all). Unbalanced input. It uses a pair of BJT outputs per channel each in a big plastic case. It makes about 150W per channel into 4 ohms and about 100W per at 8 Ohms...I'm a little fuzzy on the 8 Ohm ouput at the moment. I think its rated at 75W per ch at 8 Ohms? Seems like typical class A/B design, though I've not seen nor made a schematic...yet!

The first thing to go were those crappy push terminals for the speaker outs. Buh-bye! Used a jigsaw to open up the rectangular hole and installed banana jacks mounted on a piece of unplated fiberglass PCB. Input jacks (1/4 inch) and the 1/4 inch output jacks are also mounted on the sub PCB with the push terms. I removed the whole I/O PCB and then desoldered the plastic 1/4 inch input jacks to reuse. I liked the idea of not grounding the inputs to the chassis directly and this has worked fine.

More later! Peace!
Charlie

 

[HiString]

Charlie,

If it is OK, can we include the RA-150 series in this thread?

I will attempt to get some pics and info together over the next day or so then find someone or somewhere to host them...............just have to reinstall the camera software, etc., and then get the amp out of the rack, (something I've been putting of doing), so I can take some pics.

On one channel a hum appears as the volume knob gets around the 2 o'clock position and increases as the volume is increased............sounds like an earth type problem so this will be an opportunity to fix that if possible.

Chris

 

[Bauman]

Interested on this. I own one and it'd be great if I can improve that in any way.

cheers!
Fabio

 

[SonsOfThunder]

Here's a rundown on some of the components used in this thing and the mods so far. Locator numbers are given for reference, the same numbers apply for each channel. I assume all RA-100s are the same, but you'd better check yours before you jump in, I'm not responsible if you make your amp into a useless pile of junk.:green:

Be WARNED! PCB is single-sided and rather fragile. Most device leads are folded over, but don't be tempted to pry on the leads with an iron...you will lift them (voice of experience, if you know what I'm saying!) If you are careful you should have minimal lifting of the pads/traces. Heat the pad, then use a fine pair of pliers to lift the lead, then remove the solder with a sucker or braid.

I don't remember the stock values in some locations, perhaps someone would be so kind as to let me know what value is present in their stock unit so that we can complete the documentation?

Outputs (Q201/207) are 2SA1494/2SC3858: 200 Volt, 17 Amp, 200 Watt devices, ft=20MHz, MT200 Package (biggest I've seen at 36mm x21mm x6mm) IIRC, they are made by Sanken. Can't remember where I got data sheets but they are on the web. Driver transistors (Q205/203) are 2SC4352/2SA1668: 200V, 2A, 25W devices, ft=20MHz, FM20 package (Insulated TO-220)

The rails are +/- 50.5Vdc at idle, I measured mine stock and got 144W per channel with 4 Ohm dummy loads @ 0.2% THD, both channels loaded. That measurement was taken at 1kHz, but I got within a couple of Watts at 40Hz, 0.2% as well. I was pleasantly surprised at the amount of output for this seemingly small amp.

After modding the input/output connections, I was disappointed to find that the freq response was so poor (-1.5dB at 20k, it sounded rather cloudy and dark) with about a 10dB per octave rolloff from there upwards. Low frequency was better with only 1dB of loss at 10Hz, maybe a half dB at 20Hz. Signal-to-noise at maximum gain setting was 98.7dB on one channel, 103.7dB on the other with a 100W /4 Ohm reference. So THD and noise could both use some improvement as well as the upper freq response.

I tried a few things to increase the bandwidth and finally found the two killers by trial and substitution: C202 and 204 and the output inductor! The inductor was about 18 turns of enamelled wire on a ferrite bar core. Not wanting to just remove it entirely, I duplicated the number of turns with a same sized enam. wire with no core (or air core, if you prefer). This one was probably the biggest improvement, but I also had to change these two compensation caps (C202/204) to 101J (100pF) ceramics. Can't remember where this got me but it was now flat well past 40kHz. I'll have to measure again. I also replaced the electrolytic input cap, C215, with a pair of 1uF box film caps in parallel, one on top, one on the bottom of the PCB. I have also reduced the overall gain of the amp by lowering the feedback resistors (series and shunt) which are not marked but are on either side of C214. I'll have to measure them soon as I don't recall the paralleled values now.

The top end was nice and open now...much more listenable than stock! And the low end seemed tighter as well.

The next set of changes centered on a check of the bias adjustment. It needed a little tweek as seen on the THD sweeps to come. C203 was only 1uf...too small! A quick call to one of my workmates for a check on an appropriate value (thanks Jeri!) and a 100uF/25V low ESR cap was in place, bypassed with a 0.1uf box film for the tweekhead in me!

Then I turned to the rail caps. First, there are two 10 000uF/63V Nichicon LQ-series caps on a sub PCB that has the fuse and the connections for the power trafo and rectifier. Probably adequate capacitance for the whole amp, but the amp PCBs are connected to this board by maybe 6 inch long wires... Anyway, the only rail caps on the amp PCB are a pair of 10uf/100V...HAHAHAHA! Out those came, in went a pair of 470uF/63V Samsung FC series (what I had on hand).

I plugged and screwed it all back together and adjusted the bias and took it home to listen again. I really haven't had much time to listen to this last round of mods much, but I think its a good improvement, not as big as the first round, but still good.

Let me know if you have questions!

Peace!
Charlie

 

[Flatpicker]

Did you happen to measure the "before" and "after" slew rate? I've heard that a poor slew rate is the Achilles heel of this amp. I suppose that was indicated by the poor frequency response.

 

[SonsOfThunder]

[quote author="Flatpicker"]Did you happen to measure the "before" and "after" slew rate? I suppose that was indicated by the poor frequency response.[/quote]I'd think that listening impressions would be influenced by that output inductor and over-compensation (C202/204). (could one of you tell us what is there, stock?) I think both were in place to prevent oscillation. And I think the company figured out that this hurt thr amp bad enough that they made changes to the 150 and 200. Dunno haven't peeped one yet.

When an amp has no or not enough compensation there will be "burst oscillation" on the waveform, just before clipping...too much and the slew rate and freq response suffer.

The output inductor would limit freq response but would make the amp sound "slow" by virtue of poor damping factor. We can certainly look at some square waves, though. Just to see! There are several things I have not measured "after". I have been using this in my studio almost constant since the first mods.

Peace!
Charlie

 

[AP]

I was disappointed to find that the freq response was so poor (-1.5dB at 20k, it sounded rather cloudy and dark)

Ok, I worried when I read this, as I use an RA300 on my JBL 4312s.

I stuck the Amp on my Audio Precision (funny, I never even thought to do that before), this is what I got....

Frequency response ref 1kHz,
20kHz -0.6dB (not brilliant, but ok-ish)
10kHz -0.18dB
20Hz -0.54dB (a damn site better than my speakers...)

THD+N right channel, (see note re: right channel below) at 10V o/p (equiv 12.5Watts 8ohms) 10Hz to 80kHz is below 0.011% (mostly below 0.01%) up to 5kHz, creeping up 0.0176% at 20kHz (again, not brilliant, but competent).

Left channel noise, input pot down, -76dBV (10Hz-80kHz), -88 dBV A-wtd
Right channel noise, input pot down, -57dBV, -71dBv A-wtd

Clearly there is an issue with the Right channel noise, which no doubt is part of the THD+N reading. I didn't check yet whether it is all line frequency noise - but that is what I expect it to be... In all probability, it would turn out to be a matter of where/how the various grounds are organised (although it could be all induced hum).

... but in fact, I don't hear the noise at all ... (maybe if my PC fan was a LOT quieter.....)

The inductor was about 18 turns of enamelled wire on a ferrite bar core.

I haven't lifted the panels on mine, but I am surprised to hear that there is a ferrite cored output inductor in there. Inductors are put there for a good reason, to avoid stability problems driving capactive loads, and are usually in the range 2-10uH. However, they are always air-cored - ferrite is frowned upon as it is entirely possible to introduce non-linearity due to core saturation.

The first thing to go were those crappy push terminals for the speaker outs.

...sfunny - mine already has combo banana/binding post outputs - may be a later model (about 3 months old).

Alan

 

[AP]

Ok - couldn't resist it - took the covers off...

Looks like a fair bit has changed.

The o/p inductor is air-cored.

It looks like a discrete design (on cursory exam) - there is an op-amp in there, but that could simply be input balancing, or a DC servo (or both).

The transformer is a bit closer to the RH channel than the left. It is a laminated type, but there is a screen round it.

I can't believe how small physically the 10,000uF 63V PSU caps are. Couldn't get them anywhere near that small when I built Amps... (Ahhh! Nostalgia isn't what it used to be...)

I'm gonna put it away now, in case I get tempted to 'fix' something that isn't broken. (I simply can't be bothered...)

Alan

 

[SonsOfThunder]

RA100 was the first amp they had. Then they came out with the RA150/300/500 family as a replacement.

RA150/300/500 series are all balanced input -->RA100 is unbal
RA150/300/500 series claim DC coupling -->RA100 is not
RA150/300/500 series claim protection from faults -->RA100 does not...maybe?
RA150/300/500 series has ability to bridge -->RA100 does not
RA300/500 have LED meters
Any one of those items might require opamps.

Relay controlled on/off...is that in the output or input path? Seems like output path from the description...crappy relays could ruin damping factor...if they are indeed crappy.

For RA300 they claim 70kHz for the BW (-3dB) and 105dB SNR (22Hz ? to 22kHz BW)referenced to rated output...don't know if I have ever seen SNR spec'd as dBV before. I think they only spec'd the RA-100 to 20kHz...

Thanks for the info on the RA300, Mr. AP!

 

[AP]

don't know if I have ever seen SNR spec'd as dBV before.

Sorry - they weren't specs, they were measurments of my particular Amp.

If I had been less lazy, I have compared the noise level to max output level, to give a dB ratio to max output (the usual way), instead of against 1V.

Thanks for the history of Alesis Amps...

Rgrds
Alan

 

[SonsOfThunder]

Alan,

I knew your measurements were measurements, I guess I should have said, "I've never seen anyone measure SNR in dBV". I suppose you could spec it anyway you wanted, but many people will not know the reference to dBV, and since SNR is normally ref'd to rated power, your numbers "look" low, though they are good.

Thanks for the history of Alesis Amps...

Pssht, I'm just looking at the owners manuals on their site. :grin: I had compared Gen 1 amp versus Gen 2 amp when I first got mine. I think its encouraging to know that they listened to the feedback from their customers when they did the redesign. I do like the RA-100 for its simplicity, though.

Peace!
Charlie

 

[chingon]

Old thread but i have one of these amps and was curious if anyone has done this mod themselves??

 

[SonsOfThunder]

oops, I never posted any pics... :sad:

 

[chingon]

that would have been nice

 

[SonsOfThunder]

Alrighty Mr. chingon, (and anyone else who might be curious...)

Here are some internal pics of the RA-100 mods.

Back Panel modified for banana jack outputs, eliminate the useless 1/4 inch o/p jacks, attenuator pots moved to the back to shorten the wires and make for a simplified front panel, yet to be completed.

Back Panel Inside View shows the fiberglass board that the bananas are mounted to.

Back Panel Inside View 2 shows the input jacks (the stock ones removed from the original connector PCB, reused because they are isolated from the chassis) and the input attenuator pots.

Channel PCB Right Side shows one modded rail cap and the rewound output inductor.

Channel PCB center shows the larger electro cap on the bias (dead center near the heatsink), the modded compensation caps on either side. In the foreground you can kinda see one of the two film input caps.

Channel PCB Left Side shows the 2nd rail cap that was replaced and both film input caps. The 2nd one is peeking out from underneath!

Channel PCB Bottom View reveals the location of the film bypass on the bias electro and the second input cap. This photo is actually the PCB of the opposite side of the amp than what is shown for all the other PCB pics. This explains the different orientation of this input cap.

I have looked for the original test data and graphs, but so far have been unable to locate it amongst my "piling" system.

Any questions? :thumb:
Charlie

 

[chingon]

SWEET :thumb: Thanks Mr. Thunder. I'll have some questions when I get cracking on this new project. My question is how you did what you did with the inducter. I don't know what that is or how to do it.

 

[SonsOfThunder]

Just look at the wire that is used, estimate the size and get some in roughly the same gauge enameled magnet wire. Then get a screwdriver or something round that is approximately the same diameter as the old core. Count the number of turns on the old core and wind that same umber of turns of wire on the screwdriver. Remove the screwdriver, dress the leads and solder it in the PCB.

Clear as mud? :green:

 

[dissonantstring]

hi charlie,
so how do you go about adjusting the bias after the mods?
thanks!
-grant

 

[SonsOfThunder]

Well, i can't find any of my measurements. :evil: I'll have to bring the amp back in to work to run it on the AP.

As I recall :!: The amp had a problem with bias stability and the THD kinda wavered around until I put that larger cap there. I need to buy another one of these/get the original schematic. But you can believe that the addition of that electro cap on the bias will make a difference in low level noise and THD. The film cap might not be so noticeable, but we have those laying about here at work so its no big deal to stick it there for me.

To adjust bias, I like to use this method: Turn on a low level (<100mV) 10kHz signal into the amp with the lowest impedance load you intend to use with the amp. Look at the output on a scope. Lower the signal amplitude until you see cross-notch or until the signal disappears into the noise. If you can see cross-notch, adjust the bias just until it disappears. That should be all you need. You can adjust the bias higher using THD sweeps to determine a higher bias point, but I don't think it makes so significant of a difference in the way the amp sounds. I just aim to get rid of the cross-notch distortion and that's enough. More and you risk getting to outputs too hot and blowing up.

My amp runs cool to the touch, btw and sounds fine to me. It isn't a Levinson, but it sounds really good to my ears and since I paid about $75 plus maybe another $30 in parts, I'm very pleased. It is MUCH better than a stock RA-100 which is unusable IMO due to the upper end rolloff.

HTH!
Charlie

 

[dissonantstring]

thanks for the bias adjust tutorial charlie!:thumb:
much appreciated.
-grant