Yahoo Groups archive
Thread
2009-12-15 by Louis van Dompselaar
Does anyone have any experience with maxing out the EMAX II with DRAM from different manufacturers? I have a hard time finding 32 identical 44256, but I can get a batch of 40 different ones (which only includes a maximum of 13 from a single manufacturer). I considering getting those, unless someone already has bad experiences with mixing DRAM chips. Louis
2009-12-15 by thenewyorkcowboy
I think the important thing is the suffix at the end. During the time of competition, there were only 3 recognized timing groups, 60ns,70ns, and 80ns. I am pretty sure those are 'minimum' numbers, so for example, if the Emax is looking for a 70ns timing, a 60ns SHOULD work as well, but an 80ns would not because it is too slow. So as long as the chips have the suffix -7 or -6 you should be fine, but -8 I would expect errors. --- In emax@yahoogroups.com, Louis van Dompselaar <louis@...> wrote:
> > Does anyone have any experience with maxing out the EMAX II with DRAM > from different manufacturers? I have a hard time finding 32 identical 44256, but > I can get a batch of 40 different ones (which only includes a maximum > of 13 from a single manufacturer). > > I considering getting those, unless someone already has bad experiences > with mixing DRAM chips. > > Louis >
2009-12-16 by dwv1957
I have chips from 4 different manufacturers, and 3 different speeds, with no problems, just be aware that the system will run at whatever speed the slowest chips are. Dave --- In emax@yahoogroups.com, Louis van Dompselaar <louis@...> wrote:
> > Does anyone have any experience with maxing out the EMAX II with DRAM > from different manufacturers? I have a hard time finding 32 identical 44256, but > I can get a batch of 40 different ones (which only includes a maximum > of 13 from a single manufacturer). > > I considering getting those, unless someone already has bad experiences > with mixing DRAM chips. > > Louis >
2009-12-16 by thenewyorkcowboy
That is good to know! Since the processor is only 6Mhz there may not be that much difference of a 10ns spread. I wouldn't try any 120ns chips though, as that is nearly double. I know from PC experience that 66Mhz are 10ns, and the 100Mhz chips are 8ns, a difference of only 2ns to be able to be stable and pushed at 100mhz but we are talking about large quantities of memory compared to 8mb of an Emax running at a paltry 6Mhz. I could almost always clock a 66Mhz chip up to 83Mhz. Good luck! I learned something new too! --- In emax@yahoogroups.com, "dwv1957" <dwv1957@...> wrote:
> > I have chips from 4 different manufacturers, and 3 different speeds, with no problems, just be aware that the system will run at whatever speed the slowest chips are. > > Dave > > --- In emax@yahoogroups.com, Louis van Dompselaar <louis@> wrote: > > > > Does anyone have any experience with maxing out the EMAX II with DRAM > > from different manufacturers? I have a hard time finding 32 identical 44256, but > > I can get a batch of 40 different ones (which only includes a maximum > > of 13 from a single manufacturer). > > > > I considering getting those, unless someone already has bad experiences > > with mixing DRAM chips. > > > > Louis > > >
2009-12-16 by Louis van Dompselaar
On removing the memory expansion board I found that the original bank beneath that actually already is using two manufacturers and both 70ns and 80ns chips. So I'm probably fine using anything I can get... Op 15 dec 2009, om 20:40 heeft thenewyorkcowboy het volgende geschreven: > I think the important thing is the suffix at the end. During the time of competition, there were only 3 recognized timing groups, 60ns,70ns, and 80ns. I am pretty sure those are 'minimum' numbers, so for example, if the Emax is looking for a 70ns timing, a 60ns SHOULD work as well, but an 80ns would not because it is too slow. > > So as long as the chips have the suffix -7 or -6 you should be fine, but -8 I would expect errors. [Non-text portions of this message have been removed]
2009-12-16 by tu@...
Which processor are you referring to? The sample RAM is connected to the G chip and not the main 32CG16 CPU, which incidentally runs at 20MHz. The G chip is not documented in the Emax II service manual but I think it is likely it runs with a fixed DRAM bus cycle time in order to process the 32 voices and pass the channel data to the H chip at the fixed internal 39k sampling rate. The 44256 DRAMs are asynchronous DRAMs and are not accessed the same way as the synchronous DRAMs (SDRAM/DDR) you are familiar with on your PC. The speed grading on the 44256 DRAMs indicates the shortest access time that is guaranteed for either row or column access, often the random read/write access time will be a little slower. The chips of a given speed rating may respond faster but it is not guaranteed and may or may not work reliably at the high speed with elevated temperature etc. For SDRAM the speed is rated to the maximum speed that data can be sequentially clocked to or from the memory but random read or write access is actually much slower. Therefore you should aim to use 44256 DRAMs with a speed equal to or greater than the timing on your existing sample RAM chips. You might get away with using chips that have a slower access time but you run the risk of corruption of the data in the sample RAM :P /Tristan
On Wed, Dec 16th, 2009 at 6:44 PM, thenewyorkcowboy <thenewyorkcowboy@...> wrote: > That is good to know! Since the processor is only 6Mhz there may not > be that much difference of a 10ns spread. I wouldn't try any 120ns > chips though, as that is nearly double. > > I know from PC experience that 66Mhz are 10ns, and the 100Mhz chips > are 8ns, a difference of only 2ns to be able to be stable and pushed > at 100mhz but we are talking about large quantities of memory > compared to 8mb of an Emax running at a paltry 6Mhz. I could almost > always clock a 66Mhz chip up to 83Mhz. > > Good luck! I learned something new too! > > --- In emax@yahoogroups.com, "dwv1957" <dwv1957@...> wrote: > > > > I have chips from 4 different manufacturers, and 3 different > speeds, with no problems, just be aware that the system will run at > whatever speed the slowest chips are. > > > > Dave > > > > --- In emax@yahoogroups.com, Louis van Dompselaar <louis@> wrote: > > > > > > Does anyone have any experience with maxing out the EMAX II with > DRAM > > > from different manufacturers? I have a hard time finding 32 > identical 44256, but > > > I can get a batch of 40 different ones (which only includes a > maximum > > > of 13 from a single manufacturer). > > > > > > I considering getting those, unless someone already has bad > experiences > > > with mixing DRAM chips. > > > > > > Louis > > > > > > > >
2009-12-16 by tu@...
I mean use DRAMs with a speed rated equal to or faster than what you have already. /Tristan
On Wed, Dec 16th, 2009 at 9:01 PM, tu@... wrote: > Which processor are you referring to? The sample RAM is connected to > the G chip and not the main 32CG16 CPU, which incidentally runs at > 20MHz. > The G chip is not documented in the Emax II service manual but I > think it is likely it runs with a fixed DRAM bus cycle time in order > to process the > 32 voices and pass the channel data to the H chip at the fixed > internal 39k sampling rate. > > The 44256 DRAMs are asynchronous DRAMs and are not accessed the same > way as the synchronous DRAMs (SDRAM/DDR) you are familiar with on > your PC. The speed grading on the 44256 DRAMs indicates the shortest > access time that is guaranteed for either row or column access, often > the > random read/write access time will be a little slower. The chips of a > given speed rating may respond faster but it is not guaranteed and > may or may > not work reliably at the high speed with elevated temperature etc. > For SDRAM the speed is rated to the maximum speed that data can be > sequentially clocked to or from the memory but random read or write > access is actually much slower. > > Therefore you should aim to use 44256 DRAMs with a speed equal to or > greater than the timing on your existing sample RAM chips. You might > get > away with using chips that have a slower access time but you run the > risk of corruption of the data in the sample RAM :P > > /Tristan > > On Wed, Dec 16th, 2009 at 6:44 PM, thenewyorkcowboy > <thenewyorkcowboy@...> wrote: > > > That is good to know! Since the processor is only 6Mhz there may > not > > be that much difference of a 10ns spread. I wouldn't try any > 120ns > > chips though, as that is nearly double. > > > > I know from PC experience that 66Mhz are 10ns, and the 100Mhz > chips > > are 8ns, a difference of only 2ns to be able to be stable and > pushed > > at 100mhz but we are talking about large quantities of memory > > compared to 8mb of an Emax running at a paltry 6Mhz. I could > almost > > always clock a 66Mhz chip up to 83Mhz. > > > > Good luck! I learned something new too! > > > > --- In emax@yahoogroups.com, "dwv1957" <dwv1957@...> wrote: > > > > > > I have chips from 4 different manufacturers, and 3 different > > speeds, with no problems, just be aware that the system will run > at > > whatever speed the slowest chips are. > > > > > > Dave > > > > > > --- In emax@yahoogroups.com, Louis van Dompselaar <louis@> > wrote: > > > > > > > > Does anyone have any experience with maxing out the EMAX II > with > > DRAM > > > > from different manufacturers? I have a hard time finding 32 > > identical 44256, but > > > > I can get a batch of 40 different ones (which only includes a > > maximum > > > > of 13 from a single manufacturer). > > > > > > > > I considering getting those, unless someone already has bad > > experiences > > > > with mixing DRAM chips. > > > > > > > > Louis > > > > > > > > > > > > > > > > >
2009-12-16 by Everett
I'm not sure where you're located, but Jameco is the only place I've found these stateside without a minimum order. I do not recall what speed the chips were in the EMAX upgrade, but here's the 1Mbit 80ns version. I think they also carry a 70ns version for a few more cents. http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=42315 --- In emax@yahoogroups.com, tu@... wrote:
> > I mean use DRAMs with a speed rated equal to or faster than what you have already. > > /Tristan > > On Wed, Dec 16th, 2009 at 9:01 PM, tu@... wrote: > > > Which processor are you referring to? The sample RAM is connected to > > the G chip and not the main 32CG16 CPU, which incidentally runs at > > 20MHz. > > The G chip is not documented in the Emax II service manual but I > > think it is likely it runs with a fixed DRAM bus cycle time in order > > to process the > > 32 voices and pass the channel data to the H chip at the fixed > > internal 39k sampling rate. > > > > The 44256 DRAMs are asynchronous DRAMs and are not accessed the same > > way as the synchronous DRAMs (SDRAM/DDR) you are familiar with on > > your PC. The speed grading on the 44256 DRAMs indicates the shortest > > access time that is guaranteed for either row or column access, often > > the > > random read/write access time will be a little slower. The chips of a > > given speed rating may respond faster but it is not guaranteed and > > may or may > > not work reliably at the high speed with elevated temperature etc. > > For SDRAM the speed is rated to the maximum speed that data can be > > sequentially clocked to or from the memory but random read or write > > access is actually much slower. > > > > Therefore you should aim to use 44256 DRAMs with a speed equal to or > > greater than the timing on your existing sample RAM chips. You might > > get > > away with using chips that have a slower access time but you run the > > risk of corruption of the data in the sample RAM :P > > > > /Tristan > > > > On Wed, Dec 16th, 2009 at 6:44 PM, thenewyorkcowboy > > <thenewyorkcowboy@...> wrote: > > > > > That is good to know! Since the processor is only 6Mhz there may > > not > > > be that much difference of a 10ns spread. I wouldn't try any > > 120ns > > > chips though, as that is nearly double. > > > > > > I know from PC experience that 66Mhz are 10ns, and the 100Mhz > > chips > > > are 8ns, a difference of only 2ns to be able to be stable and > > pushed > > > at 100mhz but we are talking about large quantities of memory > > > compared to 8mb of an Emax running at a paltry 6Mhz. I could > > almost > > > always clock a 66Mhz chip up to 83Mhz. > > > > > > Good luck! I learned something new too! > > > > > > --- In emax@yahoogroups.com, "dwv1957" <dwv1957@> wrote: > > > > > > > > I have chips from 4 different manufacturers, and 3 different > > > speeds, with no problems, just be aware that the system will run > > at > > > whatever speed the slowest chips are. > > > > > > > > Dave > > > > > > > > --- In emax@yahoogroups.com, Louis van Dompselaar <louis@> > > wrote: > > > > > > > > > > Does anyone have any experience with maxing out the EMAX II > > with > > > DRAM > > > > > from different manufacturers? I have a hard time finding 32 > > > identical 44256, but > > > > > I can get a batch of 40 different ones (which only includes a > > > maximum > > > > > of 13 from a single manufacturer). > > > > > > > > > > I considering getting those, unless someone already has bad > > > experiences > > > > > with mixing DRAM chips. > > > > > > > > > > Louis > > > > > > > > > > > > > > > > > > > > > > > > > > >
2009-12-17 by Louis van Dompselaar
I'm in Europe so Jameco probably isn't an option. They won't tell me shipping costs until after they've shipped anyway, so that's reasonably tricky. Good price though! Cheaper than the second hand ones I was offered, and far cheaper than the $2.25/piece pulls I saw on eBay. Op 16 dec 2009, om 17:07 heeft Everett het volgende geschreven: > I'm not sure where you're located, but Jameco is the only place I've found these stateside without a minimum order. I do not recall what speed the chips were in the EMAX upgrade, but here's the 1Mbit 80ns version. I think they also carry a 70ns version for a few more cents. > > http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=42315 > [Non-text portions of this message have been removed]