Belton BTDR-1H Digital Reverb Module

[Taylor]

We got to talking today about the BTDR-1H delay modules that he uses in his delay pedals. As it turns out, these devices are NOT bucket brigade delays as many people seem to think but are, in fact, completely digital delays.

But that's what everyone's been saying all along, eh? Doug Deeper posted that in the first couple of replies of this thread.

[~arph]

Yes, but I think he did not mean us here.. but the people that buy them

[Cliff Schecht]

We got to talking today about the BTDR-1H delay modules that he uses in his delay pedals. As it turns out, these devices are NOT bucket brigade delays as many people seem to think but are, in fact, completely digital delays.

But that's what everyone's been saying all along, eh? Doug Deeper posted that in the first couple of replies of this thread.

It's an erroneous assumption that a lot of non-techies make. IMO it doesn't matter if a delay is analog or digital, I just care if it sounds good!

[Paul Marossy]

We got to talking today about the BTDR-1H delay modules that he uses in his delay pedals. As it turns out, these devices are NOT bucket brigade delays as many people seem to think but are, in fact, completely digital delays. The datasheet is quick to point out that this is a digital device. Another thing he pointed out is that out of the 1000 or so modules he orders at a time, about 1/3rd of them are unacceptably noisy and another 1/3rd are fairly noisy, with the last bit being the useful ones that go in his pedals. He has to hand sort every device to find ones that are usable and I'm sure by now has a large collection of unusable delay modules. His big problem is that the manufacturer REFUSES to acknowledge that there is a consistency problem from device to device and won't allow the unusable devices to be returned/traded. The reason I bring this up is because I remember reading Paul Morossy's post about building a GGG project based off of the BTDR-1H and was having some noise problems with his pedal. It may not have been the op amp but, in fact, the delay module. If anybody else has had a problem with their BTDR-1's being too noisy then please let me know, Lee could use some backup in getting the manufacturer to produce a product that actually performs as specified.

My noise problems seem to have been caused by things other than the Belton reverb module. I've got that sorted out now.

One thing I'd like to point out is that the patent documents for the Belton reverb unit refer to a PT2399. I don't know if they are actually in the unit or not, but it does mention it in the document. http://www.pat2pdf.org/patents/pat20090003614.pdf

It appears to me that it could be basically multiple PT2399s + filtering + feedback + modulation.

[neunaber]

Another thing he pointed out is that out of the 1000 or so modules he orders at a time, about 1/3rd of them are unacceptably noisy and another 1/3rd are fairly noisy, with the last bit being the useful ones that go in his pedals. He has to hand sort every device to find ones that are usable and I'm sure by now has a large collection of unusable delay modules. His big problem is that the manufacturer REFUSES to acknowledge that there is a consistency problem from device to device and won't allow the unusable devices to be returned/traded. ... If anybody else has had a problem with their BTDR-1's being too noisy then please let me know, Lee could use some backup in getting the manufacturer to produce a product that actually performs as specified.

I can say for a fact that the BTDR-1 modules meet the published spec: http://www.neunabertechnology.com/BTDR1/downloads/BTDR-1Datasheet.pdf.  I know, because I wrote the spec to conform to the test results of literally hundreds of samples.  Also, Belton tests the modules at the factory to ensure they meet the spec.  The residual noise of the modules range from -72dBV worst-case, as clearly stated in the datasheet, to around -86dBV best-case.  Yes, this is inconsistent (which has always been acknowledged) but within the specification.

I designed my sample pedal schematic using a module with -72dBV noise, and it works pretty well (http://www.neunabertechnology.com/BTDR1/downloads/ReverbPedalSchematic.pdf).  To my knowledge, other pedal makers who use the module do not have a problem with it.

I don't know what's in Lee's circuit, but my guess is that he has too much headroom and/or the gain structure is wrong.  I have offered him numerous suggestions and even offered to review his design under nondisclosure agreement.  He refuses my help and refuses to consider that perhaps his circuit is the problem rather than the module.

If someone needs a reverb that has higher signal-to-noise ratio than what the BTDR-1 offers, I have no problem recommending someone else's solution.  The Spin FV-1 and Wavefront AL3201BG are viable options.  They cost more (as a system solution), are more complicated to use, and some don't care for the sound of their built-in reverbs; but, on the other hand, they have the potential to give much better results with a little effort.  You get what you pay for.

[Paul Marossy]

I personally like the BTDR-1, I think it sounds great. IMO, it is a breakthrough device. Maybe not a 100% consistent device, but I don't think it's too bad, all things considered.

[Nasse]

Any sound samples somewhere of long, short and medioum versions?

[octfrank]

Have you heard the built in effects in the FV-1?  They aren't very good, IMO.  Even SpinSemi admits they are only for demonstration purposes.  The FV-1 is only useful if you know DSP and can write custom algorithms for it.

The Belton BTDR-1 reverb is much more musical sounding and is less expensive when considering the total system cost.

While I must disagree on your opinion of the built in reverbs, they are not for demo only. Many people use them in production, the older datasheets did caution about using in production because we could only simulate the chip and programs prior to fab (and simulating a DSP running a program in Verilog takes a LONG time to run to get just a few seconds of audio!). Once the chip and programs were proven we have told people it is fine to use the built in programs if they wish in their products. Datasheet was updated to reflect this (I believe I missed one mention of this in the the datasheet, will fix soon). Costs depend on many factors, if using a built in program and purchasing in production quantities the FV-1 is very cost effective, especially if you want more than one effect in the target device. I don't know what the Belton unit goes for in production quantities. But the FV-1 and Belton unit are different animals, the Belton unit is a fixed effect unit, nothing wrong with that. The FV-1 is a multi-effect unit, can use POTs to adjust effects in real time.

[neunaber]

Have you heard the built in effects in the FV-1?  They aren't very good, IMO.  Even SpinSemi admits they are only for demonstration purposes.  The FV-1 is only useful if you know DSP and can write custom algorithms for it.

The Belton BTDR-1 reverb is much more musical sounding and is less expensive when considering the total system cost.

While I must disagree on your opinion of the built in reverbs, they are not for demo only. Many people use them in production, the older datasheets did caution about using in production because we could only simulate the chip and programs prior to fab (and simulating a DSP running a program in Verilog takes a LONG time to run to get just a few seconds of audio!). Once the chip and programs were proven we have told people it is fine to use the built in programs if they wish in their products. Datasheet was updated to reflect this (I believe I missed one mention of this in the the datasheet, will fix soon). Costs depend on many factors, if using a built in program and purchasing in production quantities the FV-1 is very cost effective, especially if you want more than one effect in the target device. I don't know what the Belton unit goes for in production quantities. But the FV-1 and Belton unit are different animals, the Belton unit is a fixed effect unit, nothing wrong with that. The FV-1 is a multi-effect unit, can use POTs to adjust effects in real time.

I stand corrected.

I also know people using the built-in programs of the FV-1 in production.  I don't care for them, but to each his own.  That said, I am using the FV-1 in my pedal (www.neunabertechnology.com/wet-reverb), albeit with my own custom reverb algorithm.  It's perfect for stuff like this.

I don't want to get into pricing, because I'm not in sales; but I do know the BTDR-1 ends up a couple dollars less than the FV-1 when considering system cost.  The FV-1 is more flexible and has better specs but has a higher learning curve.  So, there are trade-offs.  I think each has its place.

[Taylor]

Fellas, while I find this very interesting, I do think we might be venturing into uncomfortable territory with  two competitors touting the advantages of their products on this forum. I wouldn't want to see this thread closed because of a perception of spamming.

[wampcat1]

Fellas, while I find this very interesting, I do think we might be venturing into uncomfortable territory with  two competitors touting the advantages of their products on this forum. I wouldn't want to see this thread closed because of a perception of spamming.

Honestly, I don't think that would happen here... on the gear page: YES but here at diystompboxes generally Aron and the guys are more laid back - heck steve from smallbear posts here all the time about new products and I think most of us are glad he does! I know personally, I appreciate these fellows piping in the discussion.

my $.02

[Paul Marossy]

Fellas, while I find this very interesting, I do think we might be venturing into uncomfortable territory with  two competitors touting the advantages of their products on this forum. I wouldn't want to see this thread closed because of a perception of spamming.

Honestly, I don't think that would happen here... on the gear page: YES but here at diystompboxes generally Aron and the guys are more laid back - heck steve from smallbear posts here all the time about new products and I think most of us are glad he does! I know personally, I appreciate these fellows piping in the discussion.

my $.02

I agree. 

[puretube]

Have you heard the built in effects in the FV-1?  They aren't very good, IMO.  Even SpinSemi admits they are only for demonstration purposes.  The FV-1 is only useful if you know DSP and can write custom algorithms for it.

The Belton BTDR-1 reverb is much more musical sounding and is less expensive when considering the total system cost.

While I must disagree on your opinion of the built in reverbs, they are not for demo only. Many people use them in production, the older datasheets did caution about using in production because we could only simulate the chip and programs prior to fab (and simulating a DSP running a program in Verilog takes a LONG time to run to get just a few seconds of audio!). Once the chip and programs were proven we have told people it is fine to use the built in programs if they wish in their products. Datasheet was updated to reflect this (I believe I missed one mention of this in the the datasheet, will fix soon). Costs depend on many factors, if using a built in program and purchasing in production quantities the FV-1 is very cost effective, especially if you want more than one effect in the target device. I don't know what the Belton unit goes for in production quantities. But the FV-1 and Belton unit are different animals, the Belton unit is a fixed effect unit, nothing wrong with that. The FV-1 is a multi-effect unit, can use POTs to adjust effects in real time.

Thanks for clearing things up again, Frank! 

Looking forward to whatever Keith (and you guys) come up with next... 

[puretube]

Fellas, while I find this very interesting, I do think we might be venturing into uncomfortable territory with  two competitors touting the advantages of their products on this forum. I wouldn't want to see this thread closed because of a perception of spamming.

No probz, AFAIC:
the one represents a chipmaker that is of interest for all free DIY-ers,
the other represents a modulemaker (who uses various chips...) that are of interest for the same group...

no conflicts - just: WIN!

BTW: the favorite DIY-suppliers carry both...

[AnalogCustom]

Hello:
I am currently working on a digital reverb with PT2399 six modules.    
I could not download  the patent of module. Where can I get it?

Here is the link of my project with demos, schematics...   

http://sites.google.com/site/analogcustomkits/Home/reverb-digital

I heard some demos of Belton module and the sound is very similar.

Belton demos here

http://www.abbianengineering.com/misc/BeltonCarrier.html
   
(Excuse my writing. I do not speak English.)

Regards

[Paul Marossy]

Data sheet: http://www.smallbearelec.com/Projects/BTDR-1H.pdf
Patent #20090003614: http://www.pat2pdf.org/patents/pat20090003614.pdf

[MetalGuy]

I have an FV-1 based multieffect module and AL3201 based one and according to my ears /listening to the demos so far/ they both beat the c..p out of the Belton modules but that's only my opinion.
Maybe for certain cituations and designs I'll also use one of those, who knows.

I am currently working on a digital reverb with PT2399 six modules.

Common man, give me a break. Get one of the chips mentioned above and save yourself time and pain

[appliancide]

Here is a link to AnalogCustom's project page over at electro-music: http://electro-music.com/forum/viewtopic.php?t=35276

Figured I'd post a link since details are in English there. Very cool project! Thank you for taking the time to share your project and go through the hassle of translating everything to English AnalogCustom.

[SeanCostello]

The Neunaber patent looks like an analog implementation of a feedback delay network reverb, where the scattering junction is a Householder matrix. This can be viewed as equivalent (or almost equivalent except for sign) to a waveguide reverb where a single scattering junction is used, and all waveguides have equal impedence. Julius Smith first published such an idea in 1985. The patent seems to list the idea of not sending the input to all of the delay lines at once, and taking the output from one or more parts of the scattering matrix, which has been used by Miller Puckette since the mid 1980's.

The analog part of the patent seems like an "implementation" patent, which patents a certain hardware manifestation of a concept. Yamaha has a few hundred of these patents, and they are quite common in the industry. Getting an FDN running in (largely) analog hardware is a very impressive feat.

The FV-1 can certainly implement such reverbs. FDNs tend to cost a little bit more than the allpass loop reverbs on this chip, as the chip has a few instructions and a special register to enable allpass delays to be calculated in 2 instructions. However, a waveguide/Householder style matrix can be implemented on the chip in (2N+1) instructions, where N is the order of the matrix, which is a fairly reasonable cost.

I think that the built in reverbs on the FV-1 are very good, and some of the downloadable example programs on the Spin Semi website are even better. The chip has enough cycles to run Lexicon algorithms (224XL era), so it is a nice choice for reverb pedals. I have managed to get some other nice sounds out of the chip. I have spent the last year developing VST/AU plugins, but am going to revist the FV-1 for kicks.

Sean Costello
http://www.valhalladsp.com (website)
http://valhalladsp.wordpress.com (DSP nerd blog)
http://www.audiodamage.com/effects/product.php?pid=AD023 (Eos, a reverb plugin I wrote the algorithms for)

[doug deeper]

3) pt2399!
just as i has thought!
now i wish i hadnt ripped one apart!