A basic Limiter

[LeeCavanagh]

A 14V AC adaptor would be OK. The regulators will stand up to 30V in (absolute maximum), but you really don't want to drop too much voltage, because they'll get warm!

Will a crappy one add noise tho ?

[Albert H]

No. You're putting in AC from a small transformer. Any rubbish will be filtered by the DC side of the circuit - the smoothing capacitors after the rectifier diodes and the regulators will keep things stable. A simple PSU like that (as used widely in professional audio gear) is easy to build, but very effective.

[EFIALTISFM]

I used an electrolytic cap of 1u, instead of the TANTALUM version, but that should not make a difference right?

R-3M9 and 1uf TANTALUM should control release time (small/moderate delay)

Does the aluminum electrolytic cap work as well as the tantalum in the circuit? Any issues?

[Albert H]

Yes - you're not looking for long delay times, so a conventional electrolytic will usually be OK. My limiters use conventional electrolytics throughout - I've never really liked tantalum beads!

[sinus trouble]

I used an electrolytic cap of 1u, instead of the TANTALUM version, but that should not make a difference right?

R-3M9 and 1uf TANTALUM should control release time (small/moderate delay)

Does the aluminum electrolytic cap work as well as the tantalum in the circuit? Any issues?

Yes my statement may have been a bit misleading? There was a time Tantalum was considered more stable than electrolytics? But with todays technology, little or significant difference is rarely seen.

Also remember that these components can be 'user defined' They simply control how long the limiter holds active!

For example? Fast release times can cause 'Pumping' sounds! Slower release times can generally create a more 'levelled' sound but detrimental to the overall 'loudness'

So you need to get that happy medium? If that makes sense?

[EFIALTISFM]

OK. The components are labelled. I've also repositioned a couple of parts for an easier build... I usually feed the audio in through a 100k log pot.

Hi Albert,

Two questions...

1. The board labeling for the FET has it as a DGS but the 2n5457 is a DSG so is this a typographical error or should track 22 and 23 be switched the around?

2. You've mentioned feeding the audio in through a 100k log pot. Is the pot in addition to the existing 47k resistor or by replacing it?

Ta

[Albert H]

I'd recommend something more proper, with a low noise and distortion VCA such as the circuit in the following link: http://sbpcb.blogspot.it/2012/12/stereo ... s.html?m=1

There is a fundamental error in that design as shown. It's actually two mono limiters, working independently. If the limiting action isn't the same in both channels, the stereo image will be seriously messed up!

I've built limiters (and compressors) using the 2081 (and 2180) VCA chips, and the results can be exceptionally good. I added the left and right inputs together through a couple of op-amp buffers to feed a common sidechain I built a bargraph indicator to monitor the VCA control voltage, thereby showing the amount of gain reduction being applied. My sidechain used full-wave rectification of the audio (using an "ideal diode" op-amp circuit), and the time constants for attack and decay were "program dependent", making the operation of the limiter less obvious to the ear. Without using delay line methods, the basic limiter is prone to overshoots - it can (effectively) reduce the audio too much, causing a post-peak "gap" as the gain recovers to normal. This can be minimised by sensible choice of time constants, but will always be a compromise with a basic limiter, and you'll always need a post-limiter clipper to prevent over deviation on peaks.

A more sophisticated approach - without "going digital" - is to introduce a delay into the audio path, with the sidechain take-offs before the delay, and the VCAs after the delay. With the right selection of delay (roughly the same as the attack time of your limiter), you can eliminate overshoots and pumping whilst keeping really tight control of level. This is so good that it obviates the need for a clipper after the limiter to prevent sudden peaks getting through!

The classic delay line topology - as used in the 1970s Pye broadcast limiters - used capacitive and inductive "all-pass" networks, cascaded to provide a flat frequency response over the audio range. This was incredibly expensive and required dozens of custom-wound coils. My approach is to use "bucket-brigade" delay ICs, overclocked to give a very wide bandwidth and to move the clock noise (and other artefacts) well into the ultrasonic. I used MN3007 ICs (they have to be genuine Panasonic ones) clocked at about 410kHz for about 1¼ms delay, with the sidechain attack time constant set at the same..... In use, the delay is imperceptible - it's just like the monitor speakers being (roughly) 13" further from the listener! On headphones, you can't notice the delay between your speech and it appearing in your ears!

Practically speaking, the MN3007 S/N ratio is slightly over 80dB, which is not much worse than the rest of the audio chain, all noise sources considered. Its distortion is also not too bad - and much better than the distortion introduced by clipping if it's needed!

[Albert H]

OK. The components are labelled. I've also repositioned a couple of parts for an easier build... I usually feed the audio in through a 100k log pot.

Hi Albert,

Two questions...

1. The board labeling for the FET has it as a DGS but the 2n5457 is a DSG so is this a typographical error or should track 22 and 23 be switched the around?

2. You've mentioned feeding the audio in through a 100k log pot. Is the pot in addition to the existing 47k resistor or by replacing it?

Ta

The pot is additional - the wiper feeds the input resistor.

I can't remember the FETs I used to use - they were DGS pinout. Just rotate your FET so that the legs are the right way 'round!

[sinus trouble]

EFIALTIFM

I think you will be good to go with pretty much any FET? Im sure they will perform reasonable enough!

Many 'Vintage' Audio amplifiers i have repaired often needed some manipulation of driver transistor leads to get modern replacements to fit!

[EFIALTISFM]

The pot is additional - the wiper feeds the input resistor.

I can't remember the FETs I used to use - they were DGS pinout. Just rotate your FET so that the legs are the right way 'round!

Hi Albert,

Thanks for your feedback.

Just completed the kit and tested but got a problem. Audio starts and then stops and then starts and stops again and so on. All this within 3 seconds.

Any thoughts what might be causing this?

[EFIALTISFM]

EFIALTIFM

I think you will be good to go with pretty much any FET? Im sure they will perform reasonable enough!

Many 'Vintage' Audio amplifiers i have repaired often needed some manipulation of driver transistor leads to get modern replacements to fit!

Thanks sinus.
Used a 2N5457 but had to switch round the SG legs to fit the boards layout (as suggested by Albert).
The kit has been assembled but got an issue when audio starts (I've asked for Alberts advice in my previous message)

[EFIALTISFM]

Here are some photos of the kit

[radium98]

bf245/bf244 could replace bf256 ?

[rigmo]

If you take a look at http://sound.whsites.net/project67.htm to see a limiter that's nearly the same - Rod Elliott obviously went to the same school as LB! Rod's article explains the operation of the gain cell and the steps taken to minimise distortion (I think that LB pinched Rod's values when he did his PCB).

This site can’t be reached

[thewisepranker]

Rod had to move host (and therefore domain) as Westhost supposedly screwed him over. New link: https://sound-au.com/project67.htm

[LeeCavanagh]

A friend asks if there is a pcb available ? As he wants to try it. Personally I use Breakaway One which is an awesome processor for the cost.

Will you be selling kits on eBay if they are made a little more pro with nice laid out pc, maybe even nicer built versions not just kits ? kits could be £40 and built £60. ? Or anyone else better ideas?

[sinus trouble]

The problem with 'Limiters' is that there are no 'perfect settings'

Especially with Pirate radio! Content is usually taken from a vast number of sources which will have different dynamics!

A show takes a lot of pre preparation to get it right! The work you put in FIRST will produce better results!

My advice would be to process your audio media as best you can before even thinking of going LIVE! (Levelling up etc...)

The 'Limiter' should be a last resort for protection!

[mpx]

How about a raspberry pi running stereo tool

[Albert H]

Unfortunately, unless you use a Pi 4, the R Pi doesn't have enough "grunt". Stereotool is a Windoze application - the R Pi runs Linux....

I've been experimenting with a R P1 3B+ and a sound card for audio processing. The sound card I used is an "AudioInjector" (and Australian project) and used streamed audio input, using DarkIce as the audio streamer.... Results are getting good, but it's not ready for serious use yet.....

[mike123]

Hi,

I build this processor with the original 2N5459 but used the SMD version MMBF5459.
It works perfectly, so I build a second version, but I was forced to use a 2N5458 which I thought is a replacement that should work.
However, this version only amplifies the signal (compressing) but no limiting.
The orange peak led is ON all the time. It does not blink when I put audio into it.
Can this be caused by the 2N5458?

I check the print over and over again if I made any other mistakes.