FB-01—Help with ways to stop treble roll-off—What is FB-01G?

[db7]

Alright, let’s do some puzzlin’!

Although my unit hasn’t arrived yet (and why wait till then to start making huge posts about it? ), judging by the recordings that I’ve heard, I feel more and more like I’ll end up being annoyed by the noticeable roll-off the FB-01 applies to high frequencies. I’m definitely not the only person who’s noticed this issue, as I discussed in another recent post. However, I am one of very few people who seem to have ever mentioned this issue online, and even then, there’s been no talk about possible ways to alleviate it.

So, let’s get started, shall we?

Now, presuming (logically or not?) that this effect must be caused in the analogue stage after the DAC (YM3012), I started trying to do some basic background research on what kind of things one might do to counteract attenuation of treble in a pre-amp. I’ve not found much yet, but I’m sure it’s out there, and anyway, any information I do find will take a while to be digested as I’m definitely not electr(on)ically minded.

Anyhow, it’ll be necessary to consult the service manual. So I did. And now we find something very interesting and, according to Google, never discussed beyond a mention in the title of an unproductive thread with one reply:

There’s an FB-01, and then there’s an FB-01G! I vaguely remember seeing the latter mentioned somewhere, probably in the service manual too, but no one has ever specified the differences between the two, as far as I can find. Had anyone else noticed, and does anyone know anything about, this distinction? parametric, maybe?

Well, anyhow, detectives don’t stop investigating merely because they’re waiting for statements from witnesses. Back to what I was saying about the service manual, it mentions the FB-01G with regards to some specific differences, which may or may not be very relevant to the roll-off. Have a look at this excerpt from the diagram of the complete circuit:



Look at the bottom-right, which shows the circuitry of the pre-amp leading to the two outputs. And notice the note in the box just below. The FB-01G differs from the ‘vanilla’ FB-01 in that only the FB-01G has the components named LF12 and LF13 directly prior to its outputs. Perhaps there are other differences; I need to pore through the manual to check.

Now, again, I lack the background to know things like this intuitively, but I know how to Google, and I couldn’t find any relevant answer. What type of component is denoted by “LF”, and might this be relevant to the frequency response of the output? In other words, what is the extra thing that the FB-01G is doing just before its outputs? I’m not going to be too optimistic and get my hopes up that the initialism might stand for Lowpass Filter. Besides, surely not everyone who’s uploaded recordings to YouTube has the same version of the FB-01 (i.e. G or not), whereas they all seem to suffer from lacking treble to a noticeable degree. But regardless of whether LF12 and LF13 affect frequency response, I’d really appreciate someone explaining what they do do.

More generally, whether or not LF12 and LF13 are relevant to any potential efforts to decrease the roll-off, I’d love to hear from people who know their stuff about circuits, pre-amps, and so forth. Should we assume, as I did, that it’s unlikely to be the case that the DAC is somehow inherently doomed to outputting a signal with attenuated treble? If the DAC is indeed blameless, then there must be one or more components within the pre-amp that are responsible. Which parts could be responsible for the roll-off, and how might we replace them to counteract it?

I know this is a lot to take in (as seems to be the inevitable outcome of my starting a post ), but I’d love to hear people’s thoughts and suggestions. I feel this is something that should receive a lot more attention that it has, even relative to the small volume of attention the FB-01 receives in general. The attention to this issue apparently equals almost none: myself and Alex Mauer (see the post linked above, and do check out his great songs recorded with the FB-01!) have reported the reduced treble, but I seem to be the first person who’s considered doing something about it. People complain about the FB-01 being noisy, but it seems to be OK to me: it’s the frequency response I notice, and I’d want a better one.

I’d also be grateful for any information people have about the difference between the FB-01 and the FB-01G.

Both are things that I feel should be documented. At least now that I’ve made this post, anyone else like me who searches for information specific to either of these topics will be able to find one person who’s noticed it. And with any luck, we can collaborate and find more information about both topics, maybe even a modification for the treble. This site is a fantastic resource for information on many Yamaha-related subjects, so we should take any chance we can get to further understanding of the instruments and to be there for folks who end up Googling obscure topics about them.

[Miks]

The symbols labeled 'LF' are coils. I don't know the FB-01 series but I suppose they're in there for noise suppression. Maybe a little plastic housing containing the coil or maybe a little ferrite core... some others look like a resistor...

Edit: if you'll take a look to the 'last page of the electronic parts list, first row' (at the service manual) they mention a coil (part-No: VC 516200), look at the remarks! I suppose that those letters stand for the countrys the FB is meant to be sold: J=Japan, G=Germany, F=France, C=Canada, ,U=United Kingdom or USA?

[db7]

Thanks for the quick reply and useful information, Miks! If your supposition is correct, then the FB-01G may be a revision of the FB-01 with reduced noise in its outputs.

If so, I really hope the one I’ll receive early next week is a G! I wish I’d known about this possible difference before I’d ordered one. It seems like, in any situation, there’s always one more complication to discover…

Your deducation about the abbreviations representing countries seems logical.

Any ideas at all about how to stop treble being lost in the pre-amp? parametric suggested some electronic modifications for increasing the gain, so I would imagine (as a total newb to amplifiers and everything else electronic) that there would also be some kind of mod that could alter the frequency response. I’d love it if there is and it’s quite nice and simple.

Come to think of it, since the added letter is G, is there any possibility that could stand for gain and function similarly to the mod parametric described? I guess that’s not possible and I’m just demonstrating further how little I know about electronics.

[EX5_etc]

I do not know exactly what you define as treble roll-off. I can tell you that the two coils you see over there (LF12 and LF13) are indeed low pass filters. The reason why you don't see something that looks like a low pass filter is because one part is missing from the diagram and that's the "load" as we call it in circuit theory.

Imagine for simplicity that you have a resistor (read a load) connected from the "active" terminal of audio jack JK1 to the ground terminal. That resistor is connected in series with coil LF13. The coil and the resistor form what is known as a voltage divider.

The device that you connect to JK1 (amplifier or whatever) supplies that resistor/load and in general it is known as its input impedance. That device will take the sound from JK1 and use it for whatever reason. That device reads the voltage across its input impedance. So how does it work?

The FB01 applies a voltage across the series combination of LF13 and that load. Call it Vfbo1. If you do the math you will find that the voltage across the load (the input impedance of the device that is connected to the FB01) is equal to (Vfbo1*R)/(R+ZL) where R is the unknown input impedance that for simplicity we assumed it to be a resistor. This equation is not totally correct because I left out the complex numbers (a+jb) to make things easier to understand.

ZL is equal to 2*π*f*LF13, where f is the frequency of the signal and LF13 is the value of the coil. If the frequency f increases, then ZL increases and the voltage across the load decreases. Hence high frequencies do not pass through. The opposite is true about low frequencies.

Why would you need low pass filters? Because you may be using a DAC with low resolution that cannot reproduce the signal in perfect form. Certain frequencies might be distorted and they usually come out as high pitch noises. The filters can kill those noises or reduce them a lot.

Hopefully this makes sense.
DrF

[db7]

Miks:

Sorry, I wasn’t really thinking properly earlier, so I totally missed your point! If your interpretation is correct, these coils are only present in the German version of the FB-01. I knew about the name FB-01G, I saw your paragraph about other suffixes, but somehow, I didn’t make the connection. Whoops!

That’s very interesting, then. If I’m not mistaken, you live in Germany, so can you think of any reason why there would be lowpass filters added only to units in that country?

DrF:

I really appreciate the information and the effort you put into explaining it, but sadly, I understand very little of the specifics: I’m reasonably initiated in the possible side-effects of DACs, but the electronics are, sadly, largely indecipherable to me. I have a couple of friends who know their stuff, so I’ll just have to keep pestering them to explain the basics to me step-by-step!

As for my usage of “treble roll-off”, it is a bit vague, but since I haven’t measured the effect empirically, I don’t know how to be more precise with my wording. Basically: higher frequencies are notably reduced and, past some point, seem to be absent altogether.

The easiest way to hear this is to compare various videos of the FB-01 to videos of later 4-op FM synths, or, for an extreme and very relevant (to me) comparison, compare the FB-01 to the YM2612 from the Sega Mega Drive. For instance, this video of the FB-01 against this much crisper (perhaps overly so, but nice and characteristic) recording from the Mega Drive. Sadly, no one seems to have reproduced any tracks from the latter on the former, so direct comparison is not possible, but the trends are evident across all pairs of different songs. I dare say I’ll create a direct comparison of the same song shortly after receiving my FB-01.

Since these coils were most likely only present in the German version of the module (for which, again, I’m very interested to hear any potential reason), I presume that (A) I won’t have to deal with them as I bought mine from someone in Scotland and that (B) they are not solely responsible for the declining treble of the outputs, and bypassing them would not provide a (near-)raw signal. I say that because I have never found a recording of an FB-01 that does not sound neutered, i.e. sounds full-bandwidth like other FM synths or the (infamous) YM2612 of the Mega Drive (which I’m trying to ape here!), and few to none of those were likely to be German models.

In other words, I presume prior components in the circuit are functioning as LPFs. Call me silly, but I’d like to try bypassing any or most such lowpassing circuitry to see what the near-raw output from the DAC sounds like. Half the character of the YM2612 comes from aliasing and other artefacts, and, somehow, one can grow to miss them when they’re gone! And even if I weren’t trying to simulate something with plenty aliasing and whatnot of its own, I would still want a better bandwidth from the FB-01.

What do you think? Mainly, are there other aspects earlier in the circuitry that could also be dampening the high frequencies? If so, considering my main intended uses for the module (approximation of the YM2612), I personally this would be a worthwhile project to try, with the help of experienced friends. Sure, an unmodified unit will probably be bearable for the moment and for how cheap it was—so I don’t want to sound like I already hate the thing before even receiving it —but I’m interested to know whether it can be improved.

Anyway, thanks again to you both for the educated replies.

[EX5_etc]

Hi db7.

If those two coils are present in your FB01 I would suggest that you bypass them and hear the difference. Their purpose is to attenuate the high frequencies, so you get no treble basically. Yamaha would have done this if there was some high frequency noise coming out of the DAC. So, if the coils are there, try to bypass them but don't be surprised if you hear anything annoying coming out of the little unit. I do not see their values in your picture so I cannot perhaps try to simulate their response using a hypothetical input impedance for the device that follows after them. The circuit behind the coils is more important though.

IC1 and IC4 are simply buffers, operating in what we call voltage follower configuration. Their gain is one, but they have a high input impedance and a very low output impedance. They pass all the frequencies the human ear can hear.

C18 and R18 form a first order high pass filter. Same for C8 and R8. Very low frequencies won't make it through. It's a classic design. What follows however is interesting. Take a look at the active second order low pass diagram here:

http://www.electronics-tutorials.ws/fil ... ter_5.html

Then compare it to your picture. R16, R17, C14 and C16 form the active second order low pass filter around IC4. That circuit won't allow high frequencies to go through and because it's a second order filter it will attenuate them fast. The gain of that op-amp (IC4) is given by 4.7k / 2.4k and it is about 2.

You may have to work around that circuit but it's there for a reason. Yamaha knows what that reason is. They simply try to get rid of the high frequencies. I have no idea why.

DrF

[db7]

DrF, thanks so much for taking the time to investigate this and providing an excellent analysis!

My FB-01 almost certainly won’t have the coils LF12 and LF13 due to the vanishing probability that it originated in Germany. I really wonder why the German model apparently received additional filtering in the form of the added coils, above and beyond the filtering already present in other regions.

Anyway, thanks to your expertise, we know there are other, earlier components in the circuit that function as lowpass filters. Also, I just noticed that each output, regardless of region, has a coil named LF1. What is your interpretation of these: do they supplement the prior second-order filters, and would they be possible to bypass too?

As Yamaha seem to have implemented lowpass filtering intentionally, rather than the erstwhile alternative possibility that it was due simply to bad/cheap circuitry, then my guess is it was because the DAC outputs a lot of high-frequency aliasing or similar. Pre-empting this possibility, I did search earlier for possible past discussions about the output of the YM3012, but I didn’t find anything. I wonder if its raw output is actually particularly bad or if, in fact, it would be beneficial for my tastes! Certainly, one would hope it’s not so bad that they needed to implement as steep a filter as you’ve described. If a filter of some sort is needed, I hope it’s possible to compromise by replacing it with a less steep one.

There’s a lot of food for thought here. Many thanks again for taking the time to read this and help me out.

And, more generally, your replies add yet more new and highly valuable information to the site, something that can only be good for future readers and people trekking through Google on a search for rather esoteric information such as this. That’s one of the various areas in which this site excels, and I can only foresee it getting better. For my part, I try to post new information whenever I derive it, and otherwise, I ask questions and hope a helpful expert passes by! Luckily we have some like yourself here.

It’ll be fantastic if your insights and experimentation by myself, and perhaps other owners, lead to a beneficial mod that users can do to their FB-01s. We might have to add another section to Community Hardware Projects!

[Sonus]

Deleted post

[db7]

Edit: As per the final paragraph below, there’s a fair chance my references to a German model are mistaken and that the FB-01G might be the general model. No point changing everything, so see the last paragraph for the update.

My FB-01 almost certainly won’t have the coils LF12 and LF13 due to the vanishing probability that it originated in Germany. I really wonder why the German model apparently received additional filtering in the form of the added coils, above and beyond the filtering already present in other regions.

Looking closer, the list of electrical components in the service manual describes those as “EMI Coil”. Assuming that refers to electromagnetic interference, I guess maybe they (also?) function to reduce that, which I would guess is because Germany has stricter regulations about that than the other territories? Edit: Haha, you beat me to the post, Sonus! That’s what I get for taking ages to type. Thanks a lot for the info.

The German model also has sole usage of a “Cap Special Ppec”. Moreover, there are several other differences between various regions or groups thereof, presumably due to differing properties of their power supplies, and so on:


Now, since I presume U means the USA because it tends to cluster with C for Canada(?), most notably for the fuse, it cannot refer to the United Kingdom. So, I wonder whether we perhaps do get the German or the different French version… :S Because the German model seems to be more specialised, I would assume it was for that country only, so most likely, the UK got one equivalent to the French. Logical? Edit: Not if the UK had regulations equivalent to Germany at that time, of course! Gah, so the one I receive might have even more filtering due to these coils.

Son of Son of Son of Son of Edit: The official promotional advertisement distinguishes between “U.S./Canadian models” and a “General model”. The latter might be a heavy hint as to what the G in FB-01G actually stands for; so, perhaps it wasn’t specific to Germany, and this might well mean that I can expect to find the coils in my unit.

[EX5_etc]

LF12 and LF13 might be EMI filters but they are also what I said above, low pass filters, because they are in series with the impedance of the device that follows. Depending on their inductance values, the cut-off frequency might be beyond the human audible spectrum. Thus, we may be able to ignore that functionality and assume that they are simply EMI filters, as is the case for the two LF1 coils.

These days the design has changed a bit. Instead of having separate coils as in the FB01, we couple them together. Check the pictures on the following two posts that show how I used Yamaha's design to add audio-in to my PSR1000. The input coils are easy to locate on the PCB.

http://www.yamahaforums.co.uk/forum/vie ... 195#p26250

http://www.yamahaforums.co.uk/forum/vie ... 195#p26271

Having said that, we go back to the second order filter. Ten years ago when I was teaching filters at the university (in analog electronics and control systems) I could tell you within five minutes how to properly replace the two capacitors of your filter and change the cut-off frequency. You cannot simply change them. The correcct values must be there to avoid overshoot of the filter, trouble with phase shifts, etc. Now my books and software are stored away and I can not do that. If you have a friend that knows how to use the free software TINA-TI, then you can simulate the response of the filter as the resistors and capacitors change. I have done this on my older post about the Casio CDP-100 modification.

But.... There is a BIG but here. I cannot figure out from your picture if the values of the capacitors are 0.0033 microfarad or something else. Using that value and the 2.4k resistors that compose the filter I calculate a cut-off frequency of 20kHz, beyond the hearing capabilities of most humans (I can hear very high frequencies unfortunately). The filter will start to operate a little before that but still something is off. If that frequency was at 10kHz I would say that your treble rolloff is due to this reason. At 20kHz I do not think so.

It might be the DAC after all.

DrF

[db7]

Yes, the cut-off seems more like 16 kHz, a figure I base on the facts not only that the effect is easily audible but also that the manual claims a frequency response between 10 Hz and 16 kHz.

Perhaps you’re right that it’s the DAC, but apparently it runs at almost 60 KHz, so why would it have an intrinsic roll-off at 16 KHz or so? What can cause such effects? I’m not being sceptical for its own sake: I genuinely don’t have the background to know what might cause that.

If anything, with my relative basic knowledge of DACs, I would expect that, as you proposed earlier, the output of the YM3012 is natively far too trebly, rather than not trebly enough. Such is the case with the aforementioned YM2612, which is a combined synthesiser and DAC. [edit] Well, maybe not far too trebly, but the output does seem to have a noticeable crispness to it when heard raw. [/edit]

Combining all this with the fact that the second-order filter would slope off at 12 dB per octave, which does seem in line with how noticeable the lack of treble is, it seems far more plausible. So, if you have any flashes of inspiration about how to make the numbers/units make sense, let us know!

Having said all that, I did just look up some videos on YouTube of music from old arcade games that used the same operating chip (YM2151) and either the mono (YM3012) or stereo (YM3014) DACs that typically accompanied it. Examples are Marble Madness, R-Type, and Outrun. Now, I can’t say for sure whether these recordings were made from real hardware or emulators, but they admittedly sound less trebly than the MD and closer to the FB-01, although I think the latter may still have additional filtering to some degree. So, maybe your suspicion about the DAC is correct, although the possible reasons elude me. Could it be the case, alternatively, that the YM2151 itself has a dull output, even before it reaches the DAC?

Depending on what the cause really is, maybe nothing can actually be done about this, which would be a shame. But I’m hopeful that we can puzzle through it. I say we, but all I can really do is shout encouragement. Thanks again for all of your useful replies.

[db7]

I received the FB-01 on Tuesday and have been enjoying tinkering with it since, including an audition for its new role in the band, which I’m happy to say it passed. I’ve also, after a gruelling process of trial-and-error thanks to the particular nature of Yamaha’s service manuals, worked out the proper format of sysex and can now bundle voices into banks and send them from my PC.

Confirming my later speculation about what the FB-01G was, my unit has that designation, indicating that the G does indeed stand for general, not German!

The treble is maybe not quite as attenuated as I expected. It’s still easily noticeable, don’t get me wrong, but not completely unusable, at least not with a decent EQ tacked onto its outputs. For all I know, some genuine Mega Drive units might have a similar roll-off, due to the wide variation between models in the post-DAC circuitry. For live use, the FB-01 might not need EQing, but if it did, a boost at the top end should make it acceptable. I still want to get a more hi-fi 4-op, but the FB-01 is usable enough in the meantime. Just use an EQ, don’t try comparing it to other 4-ops side-by-side, and ignorance will be bliss.

And besides, the unaltered frequency response of the FB-01 might appeal to some people. For those who are more concerned about simulating the YM2151 as used in old arcade games, if the videos I’ve seen on YouTube are anything to go by, the FB-01 is pretty much right-on as those systems also seemed to roll off the treble noticeably. And, of course, separate from video games, some people might simply like the sound of the FB-01. I definitely like it in its own way, especially for fat basses, and I’ll probably keep it for that reason even if I do get a more trebly synth to replace it for my MD-style machinations.

Since modding isn’t quite so urgently needed as I thought it might be, I’ll probably let the little guy relax for a while! But no doubt I’ll start itching to do some mods eventually, so I still welcome any input anyone has on possible reasons for the roll-off and possible ways to circumvent it, for it would benefit everyone to know, and it may convince me to get my screwdriver and electronically-minded friends out after all.

I think I’ll also want to do the above-linked mod for volume that was recollected by parametric since the output of the FB-01 is indeed low and requires a lot of amplification, which introduces noise, especially in the same high-frequency ranges that I’m wanting to EQ up!

But again, it’s not unusable even from the factory and comes alive with a bit of EQ. All in all, a nice little box, with lots of character whichever way you look at it, and I personally like the sparse interface, which is pleasantly simple but effective. Quite underrated, IMHO. I might even pick up a second unit as a spare.

[db7]

I have another big post! This all happened a while back, but I never got around to finishing my write-up. Well, I’m seemingly in that kind of mood, and I believe information needs to be free, so… This post is long because I spent so much time on this issue – and it used to be larger, so be glad I spent ages trimming it down

Since only two or three people have ever posted about this, and I’m the only one who’s taken it right to the source, I think I can lay claim to the solution. So, I’d appreciate/expect acknowledgement if anyone else applies my research. Whilst I do all this work for my own amusement, knowing someone appreciates it is what’ll keep me sharing my results with the internet. Thanks are due to DrF and others in this thread, and a physics-y friend of mine. Speaking of DrF, as some background:

But.... There is a BIG but here. I cannot figure out from your picture if the values of the capacitors are 0.0033 microfarad or something else. Using that value and the 2.4k resistors that compose the filter I calculate a cut-off frequency of 20kHz, beyond the hearing capabilities of most humans (I can hear very high frequencies unfortunately). The filter will start to operate a little before that but still something is off. If that frequency was at 10kHz I would say that your treble rolloff is due to this reason. At 20kHz I do not think so.

It might be the DAC after all.

I can verify that (A) DrF’s calculation is, of course, correct and (B) the DAC is not to blame. Now I’ll elaborate in detail.

I spent ages searching for and collating data from disparate and hard-to-find sources, as well as bothering the aforementioned friend with many questions about electronics, both of which, back then, were completely new to me! But, quoting myself, “to cut a long story as short as I’m capable of” , the first step was to test some of the earlier speculation.

The YM2164/YM3012 aren’t stupidly lowpassed intrinsically

DrF suggested the problem might be inherent to the YM2164 (FM operator) and/or YM3012 (DAC). But how could I test this idea? Knowing the YM2164 is near-identical to the much more popular – and actually documented – YM2151 wasn’t much help as all videos of raw YM2151s weren’t well-enough produced to answer either way.

Thankfully, that changed, here: this video of someone who built a board with a YM2151, YM3012, ATmega, and output jack. We can easily hear they have no trouble outputting great signals with full bandwidth and no noise. I tracked down that person’s schematic, and their filters have an ideal fc of ~19 kHz.

I was delighted! The first proper, conclusive recording of a near-raw YM2164/YM2151! Only after did I realise the YM2164 is also in the DX100, DX21, and DX27, of which videos with plenty treble are readily available, which would have spared me the hassle of hunting for diamonds in the rough on YouTube!

So, I had finally to conclude that something must be amiss in the hardware and I’d have to examine it physically. You may chuckle at how long that took, but wouldn’t you have trusted everything in the official Service Manual and used that as justification to avoid cracking open your uncommonly pristine FB-01? Well, then, read on…

Egregious errors by Yamaha in their official schematic

Back to point (A): strings in the test mode suggest a ridiculous frequency response with the attenuation at 16 kHz being a huge 30 dB – a.k.a. 3.125% of full scale. Wanting to believe this world makes sense, I told myself those figures must have meant something else! But opening the poor FB-01 to check its circuit said otherwise…

Yamaha dropped the ball majorly here. The official schematic for the FB-01 is incorrect. The capacitors in the LPFs are claimed/supposed to have a value of 3.3 nF, but they do not. Their actual value is 18 nF. The test mode admits this tacitly, but the schematic is totally wrong vs. the reality on the PCB. I’ve verified this on – count them – three FB-01s.

Think that’s bad? Check out its practical result…

Implications

The second-order LPF with capacitors at 18 nF and resistors at 2.4 kOhm has a very early cutoff frequency at just ~3.6 kHz. This is well within the audible range. Being a 2-pole LPF means the cutoff is not only early, also steep: quartering the volume for each octave up.

No wonder it sounds so muffled when compared to its contemporaries! And no wonder people have commented on its low output. When you excise the majority of the audible spectrum, what’s to be expected? Finally, no wonder I got quite a bit of noise when I used EQ to rescue some of that poor treble from the brink of extinction.

The kicker? If Yamaha had used the values of capacitors that they tell us they did, the cutoff would be perfectly placed, just where DrF said, at just over 20 kHz. And as I concluded above, neither the FM chip nor the DAC have any traits that make that inadvisable.

I’ll concede this ‘default’ roll-off of treble isn’t fatal, and as we might deduce from the lack of discussion on it, many people don’t notice it and/or aren’t bothered. Which is great! I just thought we should have this information out there, as all information should be, so that those who would like to apply it can do so.

Intermission

but seriously: what happened here: did Yamaha just not have any of the right capacitors floating around that decade and decided to settle for whatever crap they found in a pile nearby?…

I mean not to belabour the point, and this is actually one of the less offensive examples, but this joins two other things I’ve investigated and reached brand-new conclusions about Yamaha being half-assed in their choice of electronic components.

Solution

I have two FB-01 units in which I have replaced the offending 18 nF capacitors with the claimed 3.3 nF ones, and all is well!

We get a bit more noticeable trebly artefacts resulting from the quantisation of the sine in the chips’ ROMs – which miiight be why they made the filters so aggressive, but then none of the other YM2164-using synths had such extreme filters – but, if you ask me, this adds a nice bit of character – and is better than having the non-artefact parts of the signal divested of almost all treble!

As I’ve said, I love the FB-01, and even unmodified, I’ll continue to defend it as a valid module and not the joke-in-a-box some people imply. But it’s so much better with its intended frequency spectrum restored.

There’s more…

but… maybe tomorrow. They really should pay me for making these posts

[Miks]

THAT'S what I call applied research! My hat's off for you!

They really should pay me for making these posts

Definitely!

[tffShtt]

I for one am waiting for more! Love my Fb01 but its treble is nothing compared to my Korg 707, also a 2164 chip. I'm definitely going to do the db7 mod...
@db7 How do you edit your fb01? The editor on sourceforge is ok but means I have to bring a PC for live tweaks (meaning in praxis, that the fb01 gets left at home).