Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › limited bass problem in Szekeres
New Posts  All Forums:Forum Nav:

limited bass problem in Szekeres

post #1 of 15
Thread Starter 
Just finished my first Szekeres.

It sounds very good – I can hear easily more open and wide treble compared to the golden phone hole on my CD player, but the major problem is that the bass is very thin, I can’t hear those low string plucks from the double-bass anymore.

So, I made a couple of changes to the original scheme – I put off the final anti-click resistor, the mosfets’ coupling caps and the diodes. The input signal coupling caps are WIMA MKP 1,2 uF and the output caps are ELKO RAUH electrolytics (old stock – made in 1986) 1000 uF/63V bypassed with WIMA MKP 3,3 uF. The volume pot is pure garbage – I’ll use it till I get some good Alps pot.

I use very old (USA made in 1982) and very reliable Power-One ±12-15V/1.5A linear power supply (just added an AC 0.47 uF MKP filter cap before the traf) and there is no ripples and buzz to the sound. I adjusted the PSU at 14V because at 15V I could easily setup my breakfast on the mosfets’ heatsinks. But now I couldn’t feel improvement in heat dissipation – the mosfets and the power resistors sill runs very HOT. The current resistors are 22 Ohms/10W but I think to change them to most powerful ones (20W or more), or to make them some special heatsinks.

I think the bass problem is in connection with my very old output electrolytics and with my junk interconnection cable, but I’m not sure... I think the PSU gives enough power to the mosfets (1.6A), the WIMA caps are OK, the circuit hasn’t oscillation... Or I must add back the power coupling caps? The other problem was that nobody sells good quality electrolytics (ELNA Cerafine etc) in my country (Bulgaria).

If anybody has already solve that limited-bass-problem in Szekeres, my ears are open for suggestions.
post #2 of 15
Hey,

I decouple my mosfets with 0.1uF ceramic. Then I feed both channel from 10,000uF+ power reservior. Decoupling capacitors should not be omitted for any cases. The handling of low end depends on large surges of power and you do not want to run out of power.

Your mosfet should not reach boiling temperature with 20 ohm current resistors. You can check this by simply placing a tiny droplet of water on the heatsink. If it rapidly boils, the heatsink is over boiling temperature.

You can see that I have put in quite a lot of work on my Szkeres amp. But I have to tell you, even with Cerafine electrolytics, my amp had loose bass. (After switching to AKG K240, I changed my electrolytics to 24uF Solen. Bass is much tighter. This upgrade will not work with low impedance headphones.)

I recommend lowering your electrolytics capacitance to 470uF. Bigger electrolytics worse it sounds according to my experience.

Cheerios,

Tomo
post #3 of 15
sounds like your amp has an oscillation problem.A sure sign is brittle highs and lack of bass,the heat is also an indication.The mosfet will eventually "eat itself" trying to reproduce ultrasonic freqs.
Try re routing the mosfet connections or a change of gate resistors may help.

what mosfet are you using ?
grounding scheme ?
pcb or hardwired ?

More input before we can nip this
post #4 of 15
Thread Starter 
Thank you Tomo, I'll try to replace the 1000 MFD caps with some Spargue 500 MFD and I'll post the results. What kind of cap (ceramic, MKP) I should use for the coupling caps? And what value will be better?

Rick, I use the IFR510s (made in Korea) - I think they're the genuine parts. I don't think the mosfets are oscillating because the highs are very close to the original, but they're much more 3D. I'm missing just the very low bass (under 50 Hz I think). The circuit is hardwired - If I succeed to produce the ultimate sound I'll make some PCBs and a good case. Till then I'll experimentig with the parts. The grounding is star-wired.
post #5 of 15
Thread Starter 
How could I solve the heat dissipation problem - to enlarge the mosfets' heatsinks, to enlarge the source resistors (their value or power) or to enlarge the R3 (mine were 200 Kohms)?
post #6 of 15
hat size R5 are you using and at what voltage ?
are you using heat sinks now ?
And decresing the size of the output cap reduces the low frequencies,more so with low impedence cans
post #7 of 15
Thread Starter 
For R5 (I think on the scheme is R4) I use 22 Ohms/10W resistors with 14V power supply.

From 12 to 15V they run very hot and the MOSFETs too. I couldn't understand what the problem is. Some people say that even at 10 Ohms the MOSFETs stay cool.

The MOSFETs are mounted on heatsinks about 4x10 cm each.
post #8 of 15
Hello,

Rick has a point. You should check gate resistor. You must have it. This is because mosfet gate has capacitace. The wire or the pattern connecting the input cap and the gate can freely oscillate. The gate resistors are there to damp that down.

When you face components which heats up unusually, as Rick mentions, you will have to worry about oscillation.

Gate resistors is marked "Rg"

Since you are in Europe, you will have to consider European manufactures. Also, you will have to consider European sources. European sector is not my speciality and I cannot help you too much. Nonetheless, I can direct you to a site to start with.

Andrea's Heaven:
http://www.geocities.com/ResearchTriangle/8231/

This place has tons of links/banners for audio component related companies. Most of these have European bases so you will have easy access.

Note there are Russian sites now. You should check those out too. Russian military parts are pretty good and now flowing into US very cheaply.

Check back with us,

Tomo

P.S. It is amazing where new Headfiers/Headwizers are poping up today. Cheers, to the success of both, HeadFi ad HeadWize.
post #9 of 15
Thread Starter 
Ops, it’s strange but I think Tomo is right – just change the output electrolytics with 500 uF Sprague and the bass went down a little bit. I have added and the additional big 6300 uF electrolytic at the output of the PSU – this extra reservoir lowered the bass a bit too (The PSU has already a 4700 uF cap after the rectifier). I have put in and the 2x100 nF ceramic power coupling caps but there wasn’t an audible change.

The next strange thing is that I couldn’t hear a sonic difference with or without the bypassing 3,3 uF film caps on the output. The bass is just the same with unbypassed 500 uF electrolytics. Could you give me an advice – to put off the bypassing film caps or to leave them in?

The MOSFETs and the current resistors still boil water drops. Yes, I have a 220 Ohm gate resistors but the heat problem remains. I could increase their value but will it resolve the problem?

Thank you Tomo for the usefull link, I'll check it out.
post #10 of 15
Hey,

Capacitors right after rectifiers do different things. They do not decouple the amplifier. Rather they decouple the regulators.

Film cap should be left in. This increase high freq responses. Electrolytics handles high's very badly. You should be able to hear slight recession at high end.

Any of your components should be "burning hot." They should be a bit too warm to touch. Get your multimeter and measure the voltage at the gate. This should be around 8~10V. Gate voltage determine the current thru the MOSFET while operating. The resistor value limits maximum current.

What I used to do is to get pretty solid aluminum chasis ad glue/bolt the R4 onto my chasis. My R4 was metal-clad resistor is Dale "RH" series resistor. This has very convenient mounting holes. You should be able to see it on the addendums.

As for my mosfet, I used 25W heatsink built for 2 transistors. You can see that these heatsinks are pretty big so I needed to support it with heat-resistant sealant-like glue. You can bolt these onto aluminum chasis as well. Just becarefull that metal tabs on IRF510 is an electrode, so you need to insulate these guys first.

My current amp is get warm enough to my my footwarmer.

Tomo
post #11 of 15
Thread Starter 
If I increase the value of the PSU capacitor to 20000-30000 uF is there any posibility to blow something in the PSU or in the Szekeres? I just put in and a small LED to the PSU's output to recharge the cap.
post #12 of 15
Yes,

So you should not wontonly increase capacitance.

You will need higher rating transformer, rectifier and of course wire. Also, regulator output is weak to discharges. There should be a rectifier that bypass the regulator during this discharge everytime you turn on and off. You should see this on PSU schematics.

Tomo
post #13 of 15
Thread Starter 
So it's better stay with 6300 uF value or I could increase it to 10000 uF maximum without any danger?
post #14 of 15
Hey,

That, you will have to find out for yourself.

I am pretty content with 10000uF. I have switch that cuts off power to this capacitor back. The remaining power, after turning off, will discharge thru the amplifier. So my amp will fade out. Discharge will not go back to PSU.

Even then, I am not into adding excessive capacitance. In time, circuit will get dirty or something and fail. In such case, too much capacitance will be very damaging. (too much means like 1,000,000uF or 1F)

Please note the criteria for building PSU changes with amplifiers. If you are building an amplifier with only opamps, you do not need to invest too much on PSU. Opamps themselves have very good regulation, PSRR. Also, the current expenditure is much lower so decoupling caps does not have to be very large. However, for Szkeres' amp, current expenditure is very very large. Which means you do need a large decoupling caps. Also, Szkeres amp has no PSRR. So your PSU must have good regulation.

So for Szkeres amp, you end up spending more money on PSU.

Cheerios,

Tomo
post #15 of 15
Thread Starter 
Tomo, thank you very much for your NICE help!

And my last question for tonight - till now I have thought that the LED decharges the capacitance in PSU filter caps and there weren't dangerous back surges? Am I right?
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › limited bass problem in Szekeres