AtariVox+ Few oddities

[+-^CrossBow^-]

Here is what I will do. I have another 7800 that just arrived to the ITC Lab yesterday to get a UAV installed. I will test this with my Avox+ and Dungeon stalker to see if I still have the same issues. (Again, every 7800 I have tried causes the Avox+ to reset when the wiz shows up). And since I will have it opened up anyway, it will make it that much easier to connect the probes.

 

You want me to try and check this at the joystick port side off the 7800 PCB or from the Avox+ connection side? In either case, I will likely tack on some temporary wires to make it easier to attach some alligator clips onto them from my probes.

 

In fact my scope is two channel, I could in theory hook probes up to both sets of pins and see what happens to both signals at the same time. And this will cut into my Intelly game playing LOL!

[RevEng]

Whichever side is easier, I guess. If it's the same, then the atarivox side would be preferred.

 

Thanks!

[+johnnywc]

Hi Mike,

 

I wish I had better news here, but I've had no success addressing the issue from the 6502 software end.

 

It looks like the RIOT pin connected to D0/Ready line doesn't get the signal change in time (or perhaps, not every time) to prevent the AtariVox buffer from getting stuffed.

 

As to why, it seems worse when there's joystick action involved (other RIOT pins being grounded). But changes in an individual's power adapter, console, and AtariVox, don't provide relief. Similarly, people who don't experience the problem, don't experience it on multiple consoles and power supplies.

 

I've tried only sending bytes every other frame, but that didn't eliminate it. I've tried adding a secondary software queue, but it was too difficult to accurately track the phoneme lengths. I've eliminated perceived "problem phrases", but that didn't eliminate the issue. I've tried tweaked baud rate from the 7800 driver end, with no relief.

 

I'd really like for someone with the issue to scope the power, ground, and RIOT pins involved, to look at the quality. A logic trace of the buffer getting stuffed would be nice too. So far nobody with the issue has a scope or necessary skills, nor do they have a UPS to rule out external power quality.

 

That's the state of the union, sad as it is.

Thanks so much for the explanation! Too bad it's bad news, but it is what it is. :/ The voice in WoW is played during gameplay and the joystick movement is frenetic so that's probably why it's susceptible to this issue like Dungeon Stalker.

 

I have modified the logic a bit to avoid phrases interrupting each other by putting in a 6 second delay minimum between phrases (is there a programmatic way to check the state of the AVox to see if it's currently playing a phrase?) and I also disabled two phrases that seemed to trigger the reset. One of the phrases was appropriately "One touch from my babies and you will explode!" and the break happens just as the Wizard is saying "explode" lol. James @ ZPH tried the new build and said he didn't experience any crashes after a quick test so hopefully this is a workaround, at least for the demo.

 

Let me know if there's anything I can do to help. James also has been very helpful with tests so I'm sure he would help too.

 

Thanks again!

John

[+Cafeman]

Hope you guys solve this. It really spoils the fun when the Vox reset happens.

[RevEng]

I'm glad you have a workaround, John. The 6 second grace time is a bit of a shame, but workable. If we can't get to the bottom of this, I'll probably impose a similar limitation on the 7800basic vox driver.

 

Thanks for the offer of help. I'll keep it in mind. You've already helped by giving us independent confirmation that this happens on the 2600. One of the questions I've been asked multiple times is "why doesn't this happen on 2600 atarivox games?". Thanks to your report (and the 2600 test I wrote) we know it can.

[+-^CrossBow^-]

I haven't started working on the 7800 yet, but I did test out Dungeon stalker on it last night. First game through everything was fine. Wizard shows up and I hear something that sounded like "Level up?". Then I heard his laugh.. and then I defeated him and heard Wizard Defeated. Then....

 

It played the startup tones and no more Vox+ could be heard until I power cycled the 7800. I then played the game again and it did the same thing. Played a 3rd time and this time is played the tones as if resetting or locking up again right as I killed the wizard, but then I was shocked when I heard it start talking again immediately when I picked up some arrows. It didn't reset or mess up at all for the rest of the game that I played? So perhaps it is my Avox+ or something with most of the 7800s I get. But again it did reset and stop working on the first game I played even on this 7800 that I just got in.

[+-^CrossBow^-]

In discussing this with Albert during PRGE this past weekend. It has been found that the original avox works fine on both 2600 and 7800 systems and doesn't have issues with dungeon stalker or the current wip of wizard of wor for the 2600.

 

So it is apparent that the added components to make it compatible for the vectrex is the cause as the actual speakjet chip is the same.

 

The plan for wizard of wor at least it's to only offer the original avox and not the plus going forwar.

[Albert]

The plan for wizard of wor at least it's to only offer the original avox and not the plus going forwar.

 

Well, that's a possibility, but I'd prefer that the issue with the AtariVox+ and these games is discovered and fixed.

 

..Al

[+SmittyB]

Ooh, that's interesting and probably good news. I suppose then it will be necessary to determine which components differ between versions and how they affect the signals.

 

Without knowing anything about the boards (and hardware not being a strong point for me) I'd guess at something like the amount of current being borderline to the chip's requirements so that under specific circumstances the speakjet is effectively hard reset, and the onboard memory left in an unknown state. That would explain why some get the start sound and others like me get garbled noises.

[+SmittyB]

A thought has just occurred to me, I wonder if changing the volume has an effect. I lowered it from its original setting and I wonder if it would also crash with the start sound if I hadn't

 

I'll test it in a few hours, I'll use my tester to create a small phrase, see if I can confuse the vox, if I can I'll lower the volume and try again with the same phrase.

[+-^CrossBow^-]

It would be excellent to know exactly what is causing it between the two variants of the AVox. That way the extra stuff could be clipped at the source and perhaps a switch added to enable or disable the + abilities? Honestly I'm okay if two separate versions of the vox were to be offered going forward. One for the Atari systems like the original AVox, and another specifically for the Vectrex for VecVox and VecVoice ability. That way I would not have to worry about switching dip switches nearly as often.

 

But as it stands, the original AVox appears to be 100% with Dungeon Stalker and Wizard of Wor especially, and from what I played of WOW this weekend...it has a TON of speech in it. More than I've heard from any other game. And it never had any issues that I heard during the show while I was in the AA section of the exhibit hall.

[RevEng]

In discussing this with Albert during PRGE this past weekend. It has been found that the original avox works fine on both 2600 and 7800 systems and doesn't have issues with dungeon stalker or the current wip of wizard of wor for the 2600.

 

So it is apparent that the added components to make it compatible for the vectrex is the cause as the actual speakjet chip is the same.

Ever since this issue came up, it was clear that many people suspected my vox driver code was the problem. I thank you for sharing your info, and lifting that black cloud.

 

@all: Like Albert, I'd still like to come to some resolution for AtariVox+. The problem still seems to be a bit mixed, since many consoles with AtariVox+ aren't exhibiting the glitch.

 

It could be unrelated, but James' (ZPH) 2600 console seems to have TIA timing quirks. RIOT is the problem with our glitch, but perhaps something environmental is giving both chips heartache.

[Thomas Jentzsch]

The usual wild guess from me: A weak power supply maybe?

[+-^CrossBow^-]

The usual wild guess from me: A weak power supply maybe?

 

Well, my 7800 has a DC-DC converter in it that outputs 1amp current with a steady +5. But the AVox+ did this before I put the DC-DC converter into it. And I've had plenty of other 7800s cross my path for AV mods that did the reset with dungeon stalker at pretty much the exact same place. I had one 7800 that seemed to not glitch out on that game, and then it did it on the second time the wizard showed up, so it was just longer in making it happen but it did eventually happen. Every 7800 I've worked on has done this both before and after being modded. So I'm more inclined to believe it is the AVox+ and something in that design causing it. Again, it was discussed that the AVox+ had issues with the wip of WOW Champ games, but an older Avox (Which is what was used during PRGE), did not. And WOW was being shown on a 2600 with a composite mod and the new controller multiplexer to allow 2 players from controller port 1 with an original Avox on port 2. So that should be a worst case scenario from a current draw perspective.

 

I believe it was also told to me that most initial testing for Dungeon Stalker had been done with original Avox units and not the + but perhaps I'm mistaken about that.

[+SmittyB]

Just tested mine again and the volume control has no noticeable effect on how likely the issue is to occur for me. I tested with earphones plugged in, at the maximum volume, and the minimum (completely inaudible) volume by triggering the issue and turning the volume back up to hear it. I used the below test data which just says "Hello World" and had to repeatedly trigger the phrase for a good few seconds (with a 5 frame delay between each) to get the problem. Was worth a shot.

000
021
120
022
075
023
003
183
131
146
164
000
022
074
147
151
145
176
255 

[Albert]

I can't see how the volume pot would have any affect on the AtariVox+, unless a higher volume caused it to draw significantly more power. I always set them to the max volume when building them.

 

..Al

[Andromeda Stardust]

I can't see how the volume pot would have any affect on the AtariVox+, unless a higher volume caused it to draw significantly more power. I always set them to the max volume when building them.

 

..Al

Word. The Atarivox has a powerful amp driving the headphone output. Enough that I can drive a 4 ohm 2.5" cube satallite speaker (from a defunct set of surround speakers by Cambridge Soundworks) at comfortable listening level for smallish to medium sized bedroom using the dual mono output and a spliced headphone cord. White and red wires twist together, goes to red or + terminal. Black or bare ground wire goes to minus or black terminal.

 

It never misbehaved on my 7800 using an extension cable, however the impedence of the cable/system combined may cause the 5v to droop slightly on the Avox end. Since the glitches normally occur during an audio event, perhaps the amp momentarily drains enough current to dip below 5v and crash or reset the pic microcontroller. It is very likely some 7800s may be more susceptible to vcc voltage dips at the controller port than others.

 

I might ask if people who have experienced the crashing effects on 7800 systems were driving an actual speaker from the Atarivox. You may want to sub out a 1/8" stereo to rca cable instead. For a/v modded Ataris, this is easy: connect the audio from the Atari to one side of the stereo a/v inputs. Connect the audio from the vox to the other. Trim the Atarivox pot with a 1/8" jeweler's flathead screwdriver if needed. For rf Ataris, you will need a discrete audio amp separate from the tv.

[+-^CrossBow^-]

Word. The Atarivox has a powerful amp driving the headphone output. Enough that I can drive a 4 ohm 2.5" cube satallite speaker (from a defunct set of surround speakers by Cambridge Soundworks) at comfortable listening level for smallish to medium sized bedroom using the dual mono output and a spliced headphone cord. White and red wires twist together, goes to red or + terminal. Black or bare ground wire goes to minus or black terminal.

 

It never misbehaved on my 7800 using an extension cable, however the impedence of the cable/system combined may cause the 5v to droop slightly on the Avox end. Since the glitches normally occur during an audio event, perhaps the amp momentarily drains enough current to dip below 5v and crash or reset the pic microcontroller. It is very likely some 7800s may be more susceptible to vcc voltage dips at the controller port than others.

 

I might ask if people who have experienced the crashing effects on 7800 systems were driving an actual speaker from the Atarivox. You may want to sub out a 1/8" stereo to rca cable instead. For a/v modded Ataris, this is easy: connect the audio from the Atari to one side of the stereo a/v inputs. Connect the audio from the vox to the other. Trim the Atarivox pot with a 1/8" jeweler's flathead screwdriver if needed. For rf Ataris, you will need a discrete audio amp separate from the tv.

 

That is how I have been doing this the whole time...

 

I use an 1/8" headphone cable from the AVox+ directly into the aux input on my receiver when I test it. On the AV modded consoles, I can use a mixer board to mix in the Avox sound with the audio output but I usually don't connect up the audio outputs when testing the Avox+ itself. And I don't use a joystick extension cable either. The only thing I've not done is to connect my scope to the avox side of the controller input to check for voltage drops. But since it is an analog scope I can't really capture such moments and would have to watch for them while playing somehow.

[Andromeda Stardust]

That is how I have been doing this the whole time...

 

I use an 1/8" headphone cable from the AVox+ directly into the aux input on my receiver when I test it. On the AV modded consoles, I can use a mixer board to mix in the Avox sound with the audio output but I usually don't connect up the audio outputs when testing the Avox+ itself. And I don't use a joystick extension cable either. The only thing I've not done is to connect my scope to the avox side of the controller input to check for voltage drops. But since it is an analog scope I can't really capture such moments and would have to watch for them while playing somehow.

I am just trying to brainstorm various possibilities. A momentary voltage dip caused by an excessive load could definitely trigger the pic micro to crash. Idon't know what style audio amp was used, but if it's a well designed integrated class ab, the amplifier should have high efficiency and should not sink significantly more current than the load. So for instance, connecting a low impedance load such as a 4 ohm speaker to the output will sink a lot more current during audio playback compared to a high impedance load such as a preamp stage on an amplifier or receiver (typically 10kohm impedance for auxiliary inputs).

 

Regardless of the source of voltage dip, a capacitor actoss vcc may help prevent dips. Most ic packages recommend a .1 uF diode as close to thd vcc/gnd inputs as possible. I would assume the Atarivox has a filter cap already present, without ever having inspected mine. Mine was among the last voxes assembled by the.golden.axe prior to falling ill in 2013; I hope he's doing okay. Presumably the only difference between those assembled by the.golden.axe and Albert are the lable having AtariAge branding.

 

Moving on, if a .1uF ceramic cap is insufficient to prevent voltage dips or transients on the Atarivox, you can increase the value. Some caveats with doing this, if the value of the filter cap is too high (greater than ~1uF electrolytic and you may have issues), hot plugging the Atarivox into the console may cause a transient voltage dip on the console vcc resulting in crashing the actual system upon insertion. I believe this is a common complaint with the 8bitdo receivers on Nintendo hardware (including fpga consoles such as Super NT) where inserting the receiver after booting crashes or resets the system. It's especially annoying if the 8bitdo battery goes kaput and you want to switch over to wired without ruining yiur progress.

 

Back on point, if you can live with not being able to hotplug the Atarivox, a higher rated filter cap may or may not solve the issue. If you want to try increasing filter capacitance to solve transient voltage issues, I would recommend alligator clips as opposed to solder. If increased filtering doesn't correct the glitching or causes other issues (like crashing the console upon insertion), you will want to remove it.

[+-^CrossBow^-]

I believe I already attempted this adding a cap to the +5 off the AVox+ with zero change. In fact I seem to recall the issue only got worse when I did that.

[RevEng]

The only thing I've not done is to connect my scope to the avox side of the controller input to check for voltage drops. But since it is an analog scope I can't really capture such moments and would have to watch for them while playing somehow.

You can use the 2600 joystick test, with which you reproduced the issue before. Perfect for hands-free testing.

[+-^CrossBow^-]

That's a good point. Will see if I can try that over the weekend.

[Richard H.]

Try this -

 

Remove the LM386 and bridge pins 3 and 5 with a wire link (see pic) and turn the volume pot fully clock-wise.

 

You will have to use powered speakers now (but I know a lot of guys use amplified PC speakers already).

Edited November 1, 2018 by Richard H.

[Albert]

That's interesting. I use a mixer with displays to combine the audio from the AtariVox and the 2600/7800 before it goes into the television or the Commodore 1702. I'm wondering how this will affect the volume of the AtariVox audio relative to the 2600 audio.

 

..Al

[+-^CrossBow^-]

Try this -

 

Remove the LM386 and bridge pins 3 and 5 with a wire link (see pic) and turn the volume pot fully clock-wise.

 

You will have to use powered speakers now (but I know a lot of guys use amplified PC speakers already).

Interesting,

 

I plan to pull out the 7800 tonight to check some of the new homebrews out I bought at PRGE and can certainly give this a try. Are you thinking the amp is drawing too much current then and by removing it we can see if that helps?

 

As for mixing in the audio, I do tend to use an AV receiver anyway and like Al, I have a small 4 channel mixer I've used for recording audio etc for my YT channel. But, this would present an issue with the Vectrex Audio Tap because it has a built in balanced mixer input for the AVox+ to combine it with the vectrex audio to a single output. But I can't adjust those individually, so removing this would require me to use a separate mixer as well instead of the built in line input on the Vectrex Audio Tap.